├── data ├── 1 │ └── test_notaligned.csv ├── 2 │ └── 1 ├── 3 │ └── 1 ├── 4 │ └── 1 ├── 5 │ └── 1 ├── 6 │ └── 1 ├── 7 │ └── 1 ├── 8 │ └── 1 ├── 9 │ └── 1 └── 10 │ └── 1 ├── work_dir ├── 1 │ ├── codes │ │ └── 1 │ ├── runs │ │ └── 1 │ ├── checkpoint │ │ └── 1 │ └── checkpoint_best │ │ └── 1 ├── 2 │ └── 1 ├── 3 │ └── 1 ├── 4 │ └── 1 ├── 5 │ └── 1 ├── 6 │ └── 1 ├── 7 │ └── 1 ├── 8 │ └── 1 ├── 9 │ └── 1 └── 10 │ └── 1 ├── function.py ├── loss.py ├── EMA.py ├── DataLoader.py ├── requirements.txt ├── coeff_func.py ├── README.md ├── main.py ├── model.py ├── test.py ├── trainnet.py └── flow.py /data/2/1: -------------------------------------------------------------------------------- 1 | 2 | -------------------------------------------------------------------------------- /data/3/1: -------------------------------------------------------------------------------- 1 | 2 | -------------------------------------------------------------------------------- /data/4/1: -------------------------------------------------------------------------------- 1 | 2 | -------------------------------------------------------------------------------- /data/5/1: -------------------------------------------------------------------------------- 1 | 2 | -------------------------------------------------------------------------------- /data/6/1: -------------------------------------------------------------------------------- 1 | 2 | -------------------------------------------------------------------------------- /data/7/1: -------------------------------------------------------------------------------- 1 | 2 | -------------------------------------------------------------------------------- /data/8/1: -------------------------------------------------------------------------------- 1 | 2 | -------------------------------------------------------------------------------- /data/9/1: -------------------------------------------------------------------------------- 1 | 2 | -------------------------------------------------------------------------------- /data/10/1: -------------------------------------------------------------------------------- 1 | 2 | -------------------------------------------------------------------------------- /work_dir/10/1: -------------------------------------------------------------------------------- 1 | 2 | -------------------------------------------------------------------------------- /work_dir/2/1: -------------------------------------------------------------------------------- 1 | 2 | -------------------------------------------------------------------------------- /work_dir/3/1: -------------------------------------------------------------------------------- 1 | 2 | -------------------------------------------------------------------------------- /work_dir/4/1: -------------------------------------------------------------------------------- 1 | 2 | -------------------------------------------------------------------------------- /work_dir/5/1: -------------------------------------------------------------------------------- 1 | 2 | -------------------------------------------------------------------------------- /work_dir/6/1: -------------------------------------------------------------------------------- 1 | 2 | -------------------------------------------------------------------------------- /work_dir/7/1: -------------------------------------------------------------------------------- 1 | 2 | -------------------------------------------------------------------------------- /work_dir/8/1: -------------------------------------------------------------------------------- 1 | 2 | -------------------------------------------------------------------------------- /work_dir/9/1: -------------------------------------------------------------------------------- 1 | 2 | -------------------------------------------------------------------------------- /work_dir/1/codes/1: -------------------------------------------------------------------------------- 1 | 2 | -------------------------------------------------------------------------------- /work_dir/1/runs/1: -------------------------------------------------------------------------------- 1 | 2 | -------------------------------------------------------------------------------- /work_dir/1/checkpoint/1: -------------------------------------------------------------------------------- 1 | 2 | -------------------------------------------------------------------------------- /work_dir/1/checkpoint_best/1: -------------------------------------------------------------------------------- 1 | 2 | -------------------------------------------------------------------------------- /function.py: -------------------------------------------------------------------------------- 1 | import shutil 2 | import random 3 | import torch 4 | import numpy as np 5 | 6 | def setup_seed(seed): 7 | torch.manual_seed(seed) 8 | torch.cuda.manual_seed_all(seed) 9 | np.random.seed(seed) 10 | random.seed(seed) 11 | torch.backends.cudnn.deterministic = True 12 | 13 | def copy_codes(trainpath1,trainpath2,trainpath3,trainpath4, path1,path2,path3,path4): 14 | shutil.copyfile(trainpath1, path1) 15 | shutil.copyfile(trainpath2, path2) 16 | shutil.copyfile(trainpath3, path3) 17 | shutil.copyfile(trainpath4, path4) 18 | -------------------------------------------------------------------------------- /loss.py: -------------------------------------------------------------------------------- 1 | import torch 2 | import numpy as np 3 | import torch.optim as optim 4 | import torch.nn as nn 5 | import torch.nn.functional as F 6 | 7 | def createLossAndOptimizer(net, learning_rate, scheduler_step, scheduler_gamma): 8 | loss = LossFunc() 9 | # optimizer = optim.Adam([{'params': net.parameters(), 'lr':learning_rate}], lr = learning_rate, weight_decay=5e-4) 10 | optimizer = optim.Adam([{'params': net.parameters(), 'lr': learning_rate}], lr=learning_rate, eps=1e-7) 11 | scheduler = optim.lr_scheduler.StepLR(optimizer, step_size=scheduler_step, gamma=scheduler_gamma) 12 | return loss, optimizer, scheduler 13 | 14 | class LossFunc(torch.nn.Module): 15 | def __init__(self): 16 | super(LossFunc, self).__init__() 17 | 18 | def mse_loss(self, score, label): 19 | score = torch.squeeze(score) 20 | return torch.mean((score - label) ** 2) 21 | 22 | def forward(self, score, label): 23 | mse = self.mse_loss(score, label) 24 | 25 | return mse 26 | -------------------------------------------------------------------------------- /EMA.py: -------------------------------------------------------------------------------- 1 | class EMA(): 2 | 3 | def __init__(self, model, decay): 4 | self.model = model 5 | self.decay = decay 6 | self.shadow = {} 7 | self.backup = {} 8 | 9 | def register(self): 10 | for name, param in self.model.named_parameters(): 11 | if param.requires_grad: 12 | self.shadow[name] = param.data.clone() 13 | 14 | def update(self): 15 | for name, param in self.model.named_parameters(): 16 | if param.requires_grad: 17 | assert name in self.shadow 18 | new_average = (1.0 - self.decay) * param.data + self.decay * self.shadow[name] 19 | self.shadow[name] = new_average.clone() 20 | 21 | def apply_shadow(self): 22 | for name, param in self.model.named_parameters(): 23 | if param.requires_grad: 24 | assert name in self.shadow 25 | self.backup[name] = param.data 26 | param.data = self.shadow[name] 27 | 28 | def restore(self): 29 | for name, param in self.model.named_parameters(): 30 | if param.requires_grad: 31 | assert name in self.backup 32 | param.data = self.backup[name] 33 | self.backup = {} 34 | -------------------------------------------------------------------------------- /DataLoader.py: -------------------------------------------------------------------------------- 1 | import os 2 | import torch 3 | import random 4 | import numpy as np 5 | from torch.utils.data import Dataset 6 | from PIL import Image 7 | from torchvision import transforms 8 | import torchvision 9 | 10 | class CD_128(Dataset): 11 | def __init__(self, jnd_info, root_dir, test=False): 12 | self.ref_name = jnd_info[:, 0] 13 | self.test_name = jnd_info[:, 1] 14 | self.root_dir = str(root_dir) 15 | self.gt = jnd_info[:, 2] 16 | self.test = test 17 | if test == False: 18 | self.trans_org = transforms.Compose([ 19 | transforms.Resize(1024), 20 | transforms.RandomRotation(3), 21 | transforms.RandomCrop(1000), 22 | transforms.Resize(768), 23 | transforms.ToTensor(), 24 | 25 | ]) 26 | else: 27 | self.trans_org = transforms.Compose([ 28 | transforms.Resize(1024), 29 | transforms.CenterCrop(1024), 30 | transforms.ToTensor(), 31 | ]) 32 | 33 | def __len__(self): 34 | return len(self.gt) 35 | 36 | def __getitem__(self, idx): 37 | gt = float(self.gt[idx]) 38 | full_address = os.path.join(self.root_dir, str(self.ref_name[idx])) 39 | ref = Image.open(full_address).convert("RGB") 40 | ref1 = self.trans_org(ref) 41 | full_address_test = os.path.join(self.root_dir, str(self.test_name[idx])) 42 | test = Image.open(full_address_test).convert("RGB") 43 | test1 = self.trans_org(test) 44 | 45 | return ref1, test1, gt 46 | -------------------------------------------------------------------------------- /requirements.txt: -------------------------------------------------------------------------------- 1 | absl-py==1.2.0 2 | accelerate==0.12.0 3 | astunparse==1.6.3 4 | attrs==21.4.0 5 | bit-diffusion==0.0.11 6 | brotlipy==0.7.0 7 | cachetools==5.2.0 8 | coverage==6.4.2 9 | cycler==0.11.0 10 | einops==0.4.1 11 | ema-pytorch==0.0.10 12 | flatbuffers==1.12 13 | fonttools==4.34.4 14 | future==0.18.2 15 | gast==0.4.0 16 | google-auth==2.9.1 17 | google-auth-oauthlib==0.4.6 18 | google-pasta==0.2.0 19 | grpcio==1.47.0 20 | h5py==3.7.0 21 | imageio==2.19.5 22 | iniconfig==1.1.1 23 | joblib==1.1.0 24 | keras==2.9.0 25 | Keras-Preprocessing==1.1.2 26 | kiwisolver==1.4.4 27 | libclang==14.0.1 28 | lmdb==1.3.0 29 | Markdown==3.4.1 30 | matplotlib==3.5.2 31 | mkl-fft==1.3.1 32 | mkl-service==2.4.0 33 | networkx==2.8.5 34 | oauthlib==3.2.0 35 | opencv-python==4.7.0.68 36 | opt-einsum==3.3.0 37 | packaging==21.3 38 | pandas==1.4.3 39 | Pillow==9.2.0 40 | pluggy==1.0.0 41 | protobuf==3.19.4 42 | psutil==5.9.1 43 | py==1.11.0 44 | pyasn1==0.4.8 45 | pyasn1-modules==0.2.8 46 | pyparsing==3.0.9 47 | pytest==7.1.2 48 | pytest-cov==3.0.0 49 | python-dateutil==2.8.2 50 | pytz==2022.1 51 | PyWavelets==1.3.0 52 | PyYAML==6.0 53 | requests-oauthlib==1.3.1 54 | rsa==4.9 55 | scikit-image==0.19.3 56 | scikit-learn==1.1.1 57 | scipy==1.8.1 58 | tensorboard==2.9.1 59 | tensorboard-data-server==0.6.1 60 | tensorboard-plugin-wit==1.8.1 61 | tensorflow==2.9.1 62 | tensorflow-estimator==2.9.0 63 | tensorflow-io-gcs-filesystem==0.26.0 64 | termcolor==1.1.0 65 | threadpoolctl==3.1.0 66 | tifffile==2022.5.4 67 | tomli==2.0.1 68 | torch==1.12.0 69 | torchaudio==0.12.0 70 | torchvision==0.13.0 71 | tqdm==4.64.0 72 | Werkzeug==2.1.2 73 | wrapt==1.14.1 74 | yacs==0.1.8 75 | yapf==0.32.0 76 | -------------------------------------------------------------------------------- /coeff_func.py: -------------------------------------------------------------------------------- 1 | from cgi import print_form 2 | import numpy as np 3 | import pandas as pd 4 | from scipy.stats.stats import pearsonr, spearmanr, kendalltau 5 | from scipy.optimize import fmin 6 | from math import sqrt 7 | from sklearn.metrics import mean_squared_error 8 | 9 | def logistic(t, x): 10 | return 0.5 - (1 / (1 + np.exp(t * x))) 11 | 12 | 13 | def fitfun(t, x): 14 | res = t[0] * (logistic(t[1], (x-t[2]))) + t[3] + t[4] * x 15 | return res 16 | 17 | 18 | def errfun(t, x, y): 19 | return np.sum(np.power(y - fitfun(t, x),2)) 20 | 21 | 22 | def fitfun_4para(t, x): 23 | res = t[0] * (logistic(t[1], (x-t[2]))) + t[3] 24 | return res 25 | 26 | 27 | def errfun_4para(t, x, y): 28 | return np.sum(np.power(y - fitfun(t, x),2)) 29 | 30 | 31 | def RMSE(y_actual, y_predicted): 32 | rmse = sqrt(mean_squared_error(y_actual, y_predicted)) 33 | return rmse 34 | 35 | 36 | def coeff_fit(Obj,y): 37 | temp = pearsonr(Obj, y) 38 | t = np.zeros(5) 39 | t[2] = np.mean(Obj) 40 | t[3] = np.mean(y) 41 | t[1] = 1/np.std(Obj) 42 | t[0] = abs(np.max(y) - np.min(y)) 43 | t[4] = -1 44 | signslope = 1 45 | if temp[1]<=0: 46 | t[0] *= -1 47 | signslope *= -1 48 | v = [t, Obj, y] 49 | tt = fmin(errfun, t, args=(Obj, y)) 50 | fit = fitfun(tt, Obj) 51 | cc = pearsonr(fit, y)[0] 52 | # print("plcc") 53 | srocc = spearmanr(fit, y).correlation 54 | # print("srcc") 55 | krocc = kendalltau(fit, y).correlation 56 | # print("krocc") 57 | rmse = RMSE( np.absolute(y), np.absolute(fit) ) 58 | # print("Rmse") 59 | return fit, cc, srocc, krocc, rmse 60 | 61 | 62 | def compute_stress(de,dv): #obj->delta E y->subjective->dV 63 | fcv = np.sum(de*de)/np.sum(de*dv) 64 | STRESS = 100*sqrt(np.sum((de-fcv*dv)*(de-fcv*dv))/(fcv*fcv*np.sum(dv*dv))) 65 | return STRESS 66 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | # Learning a Deep Color Difference Metric for Photographic Images 2 | 3 | ## Introduction 4 | This repository contains the official pytorch implementation of the paper ["Learning a Deep Color Difference Metric for Photographic Images"](https://openreview.net) by Haoyu Chen, Zhihua Wang, Yang Yang, Qilin Sun, and Kede Ma, IEEE Conference on Computer Vision and Pattern Recognition, 2023. 5 | 6 | Most well-established and widely used color difference (CD) metrics are handcrafted and subject-calibrated against uniformly colored patches, which do not generalize well to photographic images characterized by natural scene complexities. Constructing CD formulae for photographic images is still an active research topic in imaging/illumination, vision science, and color science communities. In this paper, we aim to learn a deep CD metric for photographic images with four desirable properties. First, it well aligns with the observations in vision science that color and form are linked inextricably invisual cortical processing. Second, it is a proper metric in the mathematical sense. Third, it computes accurate CDs between photographic images, differing mainly in color appearances. Fourth, it is robust to mild geometric distortions (e.g.,translation or due to parallax), which are often present in photographic images of the same scene captured by different digital cameras. We show that all these properties can be satisfied simultaneously by learning a multi-scale autoregressive normalizing flow for feature transform, followed by the Euclidean distance which is linearly proportional to the human perceptual CD. 7 | 8 | ## Prerequisites 9 | * python 3.10 10 | 11 | * pytorch 1.12.0 12 | 13 | * ``pip install -r requirements.txt`` 14 | 15 | ## Training 16 | To train the CD-Flow from scratch, execute the following command: 17 | ```bash 18 | python main.py --training_datadir path/to/the/dataset --work_path work_dir --datapath data --batch_size_train 4 19 | ``` 20 | For the [SPCD](https://ieeexplore.ieee.org/document/9897498) dataset, you can download via [Baidu Netdisk](https://pan.baidu.com/s/18bzu-qhpMW3PqLTlVdoZRQ?pwd=txeh) or [Google Drive](https://drive.google.com/drive/folders/1Wh9fcDPviZcYWqCpXvnsJux1mnZ5WkCf?usp=share_link). 21 | ## Evaluation 22 | To evaluate the STRESS, PLCC and SRCC of your checkpoints on test set, execute: 23 | ```bash 24 | python test.py --datadir path/to/the/dataset --work_path work_dir --datapath data --batch_size_test 4 25 | ``` 26 | ## Citation 27 | ``` 28 | @inproceedings{chen2023learning, 29 | title={Learning a Deep Color Difference Metric for Photographic Images}, 30 | author={Haoyu Chen, Zhihua Wang, Yang Yang, Qilin Sun, and Kede Ma}, 31 | booktitle={Conference on Computer Vision and Pattern Recognition 2023}, 32 | year={2023} 33 | } 34 | ``` 35 | -------------------------------------------------------------------------------- /main.py: -------------------------------------------------------------------------------- 1 | import torch 2 | from trainnet import trainNet 3 | import pandas as pd 4 | import argparse 5 | 6 | def parse_config(): 7 | parser = argparse.ArgumentParser() 8 | parser.add_argument("--seed", type=int, default=100) 9 | parser.add_argument("--resume_path", type=str, default=None) 10 | parser.add_argument("--learning_rate", type=float, default=1e-5) 11 | parser.add_argument("--scheduler_step", type=int, default=5) 12 | parser.add_argument("--scheduler_gamma", type=float, default=0.5) 13 | parser.add_argument("--batch_size_train", type=int, default=4) 14 | parser.add_argument("--batch_size_test", type=int, default=4) 15 | parser.add_argument("--n_epochs", type=int, default=50) 16 | 17 | parser.add_argument("--training_datadir", type=str, default='') 18 | parser.add_argument("--colorspace", type=str, default='rgb') 19 | parser.add_argument("--trainpath1", type=str, default='trainnet.py') 20 | parser.add_argument("--trainpath2", type=str, default='main.py') 21 | parser.add_argument("--trainpath3", type=str, default='model.py') 22 | parser.add_argument("--trainpath4", type=str, default='DataLoader.py') 23 | parser.add_argument("--work_path", type=str, default='work_dir') 24 | 25 | parser.add_argument("--datapath", type=str, default='data') 26 | parser.add_argument("--trainset", type=str, default='train.csv') 27 | parser.add_argument("--valset", type=str, default='val.csv') 28 | parser.add_argument("--testset", type=str, default='test.csv') 29 | parser.add_argument("--test_aligned_path", type=str, default=None) 30 | parser.add_argument("--test_notaligned_path", type=str, default=None) 31 | 32 | return parser.parse_args() 33 | 34 | 35 | if __name__ == '__main__': 36 | config = parse_config() 37 | path = config.datapath 38 | modelprediction = pd.DataFrame(columns=['no']) 39 | modelprediction_aligned = pd.DataFrame(columns=['no']) 40 | modelprediction_notaligned = pd.DataFrame(columns=['no']) 41 | work_path = config.work_path 42 | trainpath = config.trainset 43 | valpath = config.valset 44 | testpath = config.testset 45 | performance = pd.DataFrame(columns=['stress', 'plcc', 'srcc', 'stress_aligned', 'plcc_aligned', 'srcc_aligned', 'stress_notaligned', 'plcc_notaligned', 'srcc_notaligned']) 46 | torch.cuda.empty_cache() 47 | i = 0 48 | config.datapath = path+'/{}.csv'.format(i+1) 49 | config.work_path = work_path+'/{}'.format(i+1) 50 | config.trainset = path+'/{}/'.format(i+1)+trainpath 51 | config.valset = path+'/{}/'.format(i+1)+valpath 52 | config.testset = path+'/{}/'.format(i+1)+testpath 53 | config.test_aligned_path = path+'/{}/test_aligned.csv'.format(i+1) 54 | config.test_notaligned_path = path+'/{}/test_notaligned.csv'.format(i+1) 55 | dist1, y_true1, stress1, cc_v1, srocc_v1, dist2, y_true2, stress2, cc_v2, srocc_v2,\ 56 | dist3, y_true3, stress3, cc_v3, srocc_v3 = trainNet(config, i) 57 | 58 | performance.loc['{}'.format(i), 'stress'] = stress1 59 | performance.loc['{}'.format(i), 'plcc'] = cc_v1 60 | performance.loc['{}'.format(i), 'srcc'] = srocc_v1 61 | performance.loc['{}'.format(i), 'stress_aligned'] = stress2 62 | performance.loc['{}'.format(i), 'plcc_aligned'] = cc_v2 63 | performance.loc['{}'.format(i), 'srcc_aligned'] = srocc_v2 64 | performance.loc['{}'.format(i), 'stress_notaligned'] = stress3 65 | performance.loc['{}'.format(i), 'plcc_notaligned'] = cc_v3 66 | performance.loc['{}'.format(i), 'srcc_notaligned'] = srocc_v3 67 | performance.to_csv(config.work_path + '/modelperformance.csv', index=None) 68 | -------------------------------------------------------------------------------- /model.py: -------------------------------------------------------------------------------- 1 | import math 2 | import time 3 | import torch 4 | import torch.nn as nn 5 | from flow import * 6 | import os 7 | 8 | class CDFlow(nn.Module): 9 | def __init__(self): 10 | super(CDFlow, self).__init__() 11 | self.glow = Glow(3, 8, 6, affine=True, conv_lu=True) 12 | 13 | def coordinate_transform(self, x_hat, rev=False): 14 | if not rev: 15 | log_p, logdet, x_hat = self.glow(x_hat) 16 | return log_p, logdet, x_hat 17 | 18 | else: 19 | x_hat = self.glow.reverse(x_hat) 20 | 21 | return x_hat 22 | 23 | def forward(self, x, y): 24 | log_p_x, logdet_x, x_hat = self.coordinate_transform(x, rev=False) 25 | log_p_y, logdet_y, y_hat = self.coordinate_transform(y, rev=False) 26 | 27 | x_hat_1, y_hat_1 = x_hat[0].view(x_hat[0].shape[0], -1), y_hat[0].view(x_hat[0].shape[0], -1) 28 | x_hat_2, y_hat_2 = x_hat[1].view(x_hat[1].shape[0], -1), y_hat[1].view(x_hat[1].shape[0], -1) 29 | x_hat_3, y_hat_3 = x_hat[2].view(x_hat[2].shape[0], -1), y_hat[2].view(x_hat[2].shape[0], -1) 30 | x_hat_4, y_hat_4 = x_hat[3].view(x_hat[3].shape[0], -1), y_hat[3].view(x_hat[3].shape[0], -1) 31 | x_hat_5, y_hat_5 = x_hat[4].view(x_hat[4].shape[0], -1), y_hat[4].view(x_hat[4].shape[0], -1) 32 | x_hat_6, y_hat_6 = x_hat[5].view(x_hat[5].shape[0], -1), y_hat[5].view(x_hat[5].shape[0], -1) 33 | 34 | x_cat_65 = torch.cat((x_hat_6, x_hat_5), dim=1) 35 | y_cat_65 = torch.cat((y_hat_6, y_hat_5), dim=1) 36 | x_cat_654 = torch.cat((x_hat_6, x_hat_5, x_hat_4), dim=1) 37 | y_cat_654 = torch.cat((y_hat_6, y_hat_5, y_hat_4), dim=1) 38 | x_cat_6543 = torch.cat((x_hat_6, x_hat_5, x_hat_4, x_hat_3), dim=1) 39 | y_cat_6543 = torch.cat((y_hat_6, y_hat_5, y_hat_4, y_hat_3), dim=1) 40 | x_cat_65432 = torch.cat((x_hat_6, x_hat_5, x_hat_4, x_hat_3, x_hat_2), dim=1) 41 | y_cat_65432 = torch.cat((y_hat_6, y_hat_5, y_hat_4, y_hat_3, y_hat_2), dim=1) 42 | x_cat_654321 = torch.cat((x_hat_6, x_hat_5, x_hat_4, x_hat_3, x_hat_2, x_hat_1), dim=1) 43 | y_cat_654321 = torch.cat((y_hat_6, y_hat_5, y_hat_4, y_hat_3, y_hat_2, y_hat_1), dim=1) 44 | 45 | mse6 = (x_hat_6 - y_hat_6).view(x_hat_6.shape[0], -1) 46 | mse6 = mse6.unsqueeze(1) 47 | mse6 = torch.sqrt(1e-8 + torch.matmul(mse6, mse6.transpose(dim0=-2, dim1=-1))/mse6.shape[2]) 48 | mse6 = mse6.squeeze(2) 49 | 50 | mse65 = (x_cat_65 - y_cat_65).view(x_cat_65.shape[0], -1) 51 | mse65 = mse65.unsqueeze(1) 52 | mse65 = torch.sqrt(1e-8 + torch.matmul(mse65, mse65.transpose(dim0=-2, dim1=-1))/mse65.shape[2]) 53 | mse65 = mse65.squeeze(2) 54 | 55 | mse654 = (x_cat_654 - y_cat_654).view(x_cat_654.shape[0], -1) 56 | mse654 = mse654.unsqueeze(1) 57 | mse654 = torch.sqrt(1e-8 + torch.matmul(mse654, mse654.transpose(dim0=-2, dim1=-1))/mse654.shape[2]) 58 | mse654 = mse654.squeeze(2) 59 | 60 | mse6543 = (x_cat_6543 - y_cat_6543).view(x_cat_6543.shape[0], -1) 61 | mse6543 = mse6543.unsqueeze(1) 62 | mse6543 = torch.sqrt(1e-8 + torch.matmul(mse6543, mse6543.transpose(dim0=-2, dim1=-1))/mse6543.shape[2]) 63 | mse6543 = mse6543.squeeze(2) 64 | 65 | mse65432 = (x_cat_65432 - y_cat_65432).view(x_cat_65432.shape[0], -1) 66 | mse65432 = mse65432.unsqueeze(1) 67 | mse65432 = torch.sqrt(1e-8 + torch.matmul(mse65432, mse65432.transpose(dim0=-2, dim1=-1)) / mse65432.shape[2]) 68 | mse65432 = mse65432.squeeze(2) 69 | 70 | mse654321 = (x_cat_654321 - y_cat_654321).view(x_cat_654321.shape[0], -1) 71 | mse654321 = mse654321.unsqueeze(1) 72 | mse654321 = torch.sqrt(1e-8 + torch.matmul(mse654321, mse654321.transpose(dim0=-2, dim1=-1)) / mse654321.shape[2]) 73 | mse654321 = mse654321.squeeze(2) 74 | 75 | return mse654321, mse65432, mse6543, mse654, mse65, mse6, log_p_x, logdet_x, log_p_y, logdet_y 76 | 77 | -------------------------------------------------------------------------------- /test.py: -------------------------------------------------------------------------------- 1 | import time 2 | from EMA import EMA 3 | import torch 4 | from torch.utils.data import DataLoader 5 | from model import CDFlow 6 | from DataLoader import CD_128 7 | from coeff_func import * 8 | import os 9 | from loss import createLossAndOptimizer 10 | from torch.autograd import Variable 11 | import torchvision 12 | import torch.autograd as autograd 13 | from function import setup_seed, copy_codes 14 | import argparse 15 | 16 | def parse_config(): 17 | parser = argparse.ArgumentParser() 18 | parser.add_argument("--batch_size_test", type=int, default=4) 19 | parser.add_argument("--work_path", type=str, default='work_dir') 20 | parser.add_argument("--datapath", type=str, default='data') 21 | parser.add_argument("--dataset", type=str, default='') 22 | parser.add_argument("--testset", type=str, default='test.csv') 23 | parser.add_argument("--test_aligned_path", type=str, default=None) 24 | parser.add_argument("--test_notaligned_path", type=str, default=None) 25 | 26 | return parser.parse_args() 27 | 28 | def test(data_val_loader, net): 29 | dist = [] 30 | y_true = [] 31 | for i, data in enumerate(data_val_loader, 0): 32 | with torch.no_grad(): 33 | x, y, gts = data 34 | y_val = gts.numpy() 35 | x, y, gts = \ 36 | Variable(x).cuda(), \ 37 | Variable(y).cuda(), \ 38 | Variable(gts).cuda() 39 | score, _, _, _, _, _, _, _, _, _ = net(x, y) 40 | pred = (torch.squeeze(score)).cpu().detach().numpy().tolist() 41 | if isinstance(pred, list): 42 | dist.extend(pred) 43 | y_true.extend(y_val.tolist()) 44 | else: 45 | dist.append(np.array(pred)) 46 | y_true.append(y_val) 47 | 48 | dist_np = np.array(dist) 49 | y_true_np = np.array(y_true).squeeze() 50 | stress = compute_stress(dist_np, y_true_np) 51 | _, cc_v, srocc_v, krocc_v, rmse_v = coeff_fit(dist_np, y_true_np) 52 | 53 | return srocc_v, cc_v, stress, dist, y_true 54 | 55 | config = parse_config() 56 | path = config.datapath 57 | work_path = config.work_path 58 | testpath = config.testset 59 | workspace = work_path + '/{}'.format(1) 60 | testset = path + '/{}/'.format(1) + testpath 61 | test_aligned_path = path + '/{}/test_aligned.csv'.format(1) 62 | test_notaligned_path = path + '/{}/test_notaligned.csv'.format(1) 63 | datadir = config.dataset 64 | batch_size_test = config.batch_size_test 65 | 66 | test_pairs = np.genfromtxt(open(testset, encoding='UTF-8-sig'), delimiter=',', dtype=str) 67 | test_aligned_pairs = np.genfromtxt(open(test_aligned_path), delimiter=',', dtype=str) 68 | test_notaligned_pairs = np.genfromtxt(open(test_notaligned_path), delimiter=',', dtype=str) 69 | 70 | data_test = CD_128(test_pairs[:], root_dir=datadir, test=True) 71 | test_aligned = CD_128(test_aligned_pairs[:], root_dir=datadir, test=True) 72 | test_notaligned = CD_128(test_notaligned_pairs[:], root_dir=datadir, test=True) 73 | 74 | data_test_loader = DataLoader(data_test, batch_size=batch_size_test, shuffle=False, pin_memory=True, num_workers=4) 75 | data_test_aligned_loader = DataLoader(test_aligned, batch_size=batch_size_test, shuffle=False, pin_memory=True, 76 | num_workers=4) 77 | data_test_notaligned_loader = DataLoader(test_notaligned, batch_size=batch_size_test, shuffle=False, pin_memory=True, 78 | num_workers=4) 79 | 80 | print('#############################################################################') 81 | print("Testing...") 82 | print('#############################################################################') 83 | device = torch.device("cuda") 84 | pt = os.path.join(workspace, 'checkpoint_best', 'ModelParams_Best_val.pt') 85 | checkpoint = torch.load(pt) 86 | net = CDFlow().cuda() 87 | net = torch.nn.DataParallel(net).cuda() 88 | net.load_state_dict(checkpoint['state_dict']) 89 | net.eval() 90 | srocc_v1, cc_v1, stress1, dist1, y_true1 = test(data_test_loader, net) 91 | print('All: plcc{}; srcc{}; stress{}'.format(cc_v1, srocc_v1, stress1)) 92 | srocc_v2, cc_v2, stress2, dist2, y_true2 = test(data_test_aligned_loader, net) 93 | print('Pixel-wise aligned: plcc{}; srcc{}; stress{}'.format(cc_v2, srocc_v2, stress2)) 94 | srocc_v3, cc_v3, stress3, dist3, y_true3 = test(data_test_notaligned_loader, net) 95 | print('Non-Pixel-wise aligned: plcc{}; srcc{}; stress{}'.format(cc_v3, srocc_v3, stress3)) 96 | -------------------------------------------------------------------------------- /trainnet.py: -------------------------------------------------------------------------------- 1 | import time 2 | from EMA import EMA 3 | import torch 4 | from torch.utils.data import DataLoader 5 | from model import CDFlow 6 | from DataLoader import CD_128 7 | from coeff_func import * 8 | import os 9 | from loss import createLossAndOptimizer 10 | from torch.autograd import Variable 11 | import torch.autograd as autograd 12 | from function import setup_seed, copy_codes 13 | from math import log 14 | 15 | 16 | def trainNet(config, times): 17 | resume_path = config.resume_path 18 | learning_rate = config.learning_rate 19 | scheduler_step = config.scheduler_step 20 | scheduler_gamma = config.scheduler_gamma 21 | batch_size_train = config.batch_size_train 22 | batch_size_test = config.batch_size_test 23 | n_epochs = config.n_epochs 24 | training_datadir = config.training_datadir 25 | colorspace = config.colorspace 26 | trainpath1 = config.trainpath1 27 | trainpath2 = config.trainpath2 28 | trainpath3 = config.trainpath3 29 | trainpath4 = config.trainpath4 30 | workspace = config.work_path 31 | device = torch.device("cuda") 32 | # set random seed 33 | setup_seed(config.seed) 34 | if not os.path.exists(workspace): 35 | os.mkdir(workspace) 36 | if not os.path.exists(os.path.join(workspace, 'codes')): 37 | os.mkdir(os.path.join(workspace, 'codes')) 38 | if not os.path.exists(os.path.join(workspace, 'checkpoint')): 39 | os.mkdir(os.path.join(workspace, 'checkpoint')) 40 | if not os.path.exists(os.path.join(workspace, 'checkpoint_best')): 41 | os.mkdir(os.path.join(workspace, 'checkpoint_best')) 42 | copy_codes(trainpath1=trainpath1, trainpath2=trainpath2, trainpath3=trainpath3, trainpath4=trainpath4, 43 | path1=os.path.join(workspace, 'codes/trainNet.py'), path2=os.path.join(workspace, 'codes/main.py'), 44 | path3=os.path.join(workspace, 'codes/net.py'), path4=os.path.join(workspace, 'codes/DataLoader.py')) 45 | 46 | print("============ HYPERPARAMETERS ==========") 47 | print("batch_size_train and test=", batch_size_train, batch_size_test) 48 | print("epochs=", n_epochs) 49 | print('learning rate=', learning_rate) 50 | print('scheduler_step=', scheduler_step) 51 | print('scheduler_gamma=', scheduler_gamma) 52 | print('training dir=', training_datadir) 53 | print('colorspace=', colorspace) 54 | print(config.trainset) 55 | print(config.valset) 56 | print(config.testset) 57 | print(config.test_aligned_path) 58 | print(config.test_notaligned_path) 59 | train_pairs = np.genfromtxt(open(config.trainset, encoding='UTF-8-sig'), delimiter=',', dtype=str) 60 | val_pairs = np.genfromtxt(open(config.valset, encoding='UTF-8-sig'), delimiter=',', dtype=str) 61 | test_pairs = np.genfromtxt(open(config.testset, encoding='UTF-8-sig'), delimiter=',', dtype=str) 62 | 63 | test_aligned_pairs = np.genfromtxt(open(config.test_aligned_path), delimiter=',', dtype=str) 64 | test_notaligned_pairs = np.genfromtxt(open(config.test_notaligned_path), delimiter=',', dtype=str) 65 | 66 | data_train = CD_128(train_pairs[:], root_dir=training_datadir, test=False) 67 | data_val = CD_128(val_pairs[:], root_dir=training_datadir, test=True) 68 | data_test = CD_128(test_pairs[:], root_dir=training_datadir, test=True) 69 | test_aligned = CD_128(test_aligned_pairs[:], root_dir=training_datadir, test=True) 70 | test_notaligned = CD_128(test_notaligned_pairs[:], root_dir=training_datadir, test=True) 71 | 72 | net = CDFlow().to(device) 73 | net = torch.nn.DataParallel(net) 74 | net = net.to(device) 75 | loss, optimizer, scheduler = createLossAndOptimizer(net, learning_rate, scheduler_step, scheduler_gamma) 76 | 77 | data_train_loader = DataLoader(data_train, batch_size=batch_size_train, shuffle=True, 78 | pin_memory=True, num_workers=4) 79 | data_val_loader = DataLoader(data_val, batch_size=batch_size_test, shuffle=True, pin_memory=True, 80 | num_workers=4) 81 | data_test_loader = DataLoader(data_test, batch_size=batch_size_test, shuffle=False, 82 | pin_memory=True, num_workers=4) 83 | data_test_aligned_loader = DataLoader(test_aligned, batch_size=batch_size_test, shuffle=False, 84 | pin_memory=True, num_workers=4) 85 | data_test_notaligned_loader = DataLoader(test_notaligned, batch_size=batch_size_test, 86 | shuffle=False, pin_memory=True, num_workers=4) 87 | 88 | if resume_path is not None: 89 | checkpoint = torch.load(resume_path) 90 | start_epoch = checkpoint['epoch'] + 1 91 | net.load_state_dict(checkpoint['state_dict']) 92 | optimizer.load_state_dict(checkpoint['optimizer']) 93 | print('continue to train: shuffle{} epoch{} '.format(times + 1, start_epoch)) 94 | else: 95 | start_epoch = 0 96 | 97 | training_start_time = time.time() 98 | rows, columns = train_pairs.shape 99 | n_batches = rows // batch_size_train 100 | valsrcc = 0 101 | ema = EMA(net, 0.999) 102 | ema.register() 103 | autograd.set_detect_anomaly(True) 104 | for epoch in range(start_epoch, n_epochs): 105 | # initiate parameters for statistic recordings. 106 | dist = [] 107 | y_true = [] 108 | running_loss = 0.0 109 | total_train_loss = 0 110 | start_time = time.time() 111 | print_every = 20 112 | train_counter = 0 113 | net.train() 114 | print("---------------------train mode-------epoch{}--------------------------".format(epoch)) 115 | for i, data in enumerate(data_train_loader, 0): 116 | train_counter = train_counter + 1 117 | x, y, gts = data 118 | y_val = gts.numpy() 119 | x, y, gts = \ 120 | Variable(x).to(device), \ 121 | Variable(y).to(device), \ 122 | Variable(gts).to(device) 123 | optimizer.zero_grad() 124 | 125 | score, score65432, score6543, score654, score65, score6, log_p_x, logdet_x, log_p_y, logdet_y = net(x, y) 126 | 127 | logdet_x = logdet_x.mean() 128 | logdet_y = logdet_y.mean() 129 | 130 | loss_x, log_p_x, log_det_x = calc_loss(log_p_x, logdet_x, 768, 2.0 ** 5) 131 | loss_y, log_p_y, log_det_y = calc_loss(log_p_y, logdet_y, 768, 2.0 ** 5) 132 | 133 | score_loss = 10 * loss(score, gts) + loss(score65432, gts) + loss(score6543, gts) + loss(score654, gts) + loss(score65, gts) + loss(score6, gts) 134 | 135 | loss_size = 10 * score_loss + loss_x + loss_y 136 | loss_size.backward() 137 | optimizer.step() 138 | ema.update() 139 | 140 | running_loss += loss_size.item() 141 | total_train_loss += loss_size.item() 142 | 143 | pred = (torch.squeeze(score)).cpu().detach().numpy().tolist() 144 | if isinstance(pred, list): 145 | dist.extend(pred) 146 | y_true.extend(y_val.tolist()) 147 | else: 148 | dist.append(np.array(pred)) 149 | y_true.append(y_val) 150 | 151 | if (i + 1) % (print_every + 1) == 0: 152 | print("Epoch {}, {:d}% \t train_loss: {:.6f} took: {:.2f}s".format( 153 | epoch + 1, int(100 * (i + 1) / n_batches), running_loss / print_every, time.time() - start_time)) 154 | 155 | running_loss = 0.0 156 | start_time = time.time() 157 | 158 | torch.save( 159 | {"state_dict": net.state_dict(), 'epoch': epoch, 'optimizer': optimizer.state_dict(), 'times': times}, \ 160 | os.path.join(workspace, 'checkpoint', 'ModelParams_checkpoint.pt')) 161 | 162 | # Calculate correlation coefficients between the predicted values and ground truth values on training set. 163 | dist = np.array(dist).squeeze() 164 | y_true = np.array(y_true).squeeze() 165 | _, cc_v, srocc_v, krocc_v, rmse_v = coeff_fit(dist, y_true) 166 | print("Training set: PCC{:.4}, SROCC{:.4}, KROCC{:.4}, RMSE{:.4}".format(cc_v, srocc_v, krocc_v, rmse_v)) 167 | # validation 168 | # EMA 169 | ema.apply_shadow() 170 | # EMA 171 | net.eval() 172 | print("----------------------------validation mode---------------------------------") 173 | srocc_v, total_val_loss, val_counter, cc_v, krocc_v, rmse_v, stress, dist, y_true, score_val = test( 174 | data_val_loader, net, loss) 175 | # srocc_a, total_val_loss_a, val_counter_a, cc_a, krocc_a, rmse_a, stress_a, dist_a, y_true_a, score_a = test( 176 | # data_test_aligned_loader, net, loss) 177 | # srocc_na, total_val_loss_na, val_counter_na, cc_na, krocc_na, rmse_na, stress_na, dist_na, y_true_na, score_na = test( 178 | # data_test_notaligned_loader, net, loss) 179 | 180 | if srocc_v > valsrcc: 181 | valsrcc = srocc_v 182 | torch.save({"state_dict": net.state_dict()}, 183 | os.path.join(workspace, 'checkpoint_best', 'ModelParams_Best_val.pt')) 184 | print('update best model...') 185 | print("VALIDATION: PCC{:.4}, SROCC{:.4}, STRESS{:.4}, RMSE{:.4}".format(cc_v, srocc_v, stress, rmse_v)) 186 | print("loss = {:.6}".format(total_val_loss / val_counter)) 187 | # EMA 188 | ema.restore() 189 | # EMA 190 | scheduler.step() 191 | 192 | print('#############################################################################') 193 | print("Training finished, took {:.2f}s".format(time.time() - training_start_time)) 194 | pt = os.path.join(workspace, 'checkpoint_best', 'ModelParams_Best_val.pt') 195 | checkpoint = torch.load(pt) 196 | net = CDFlow().to(device) 197 | net = torch.nn.DataParallel(net).to(device) 198 | net.load_state_dict(checkpoint['state_dict']) 199 | net.eval() 200 | srocc_v1, total_val_loss, val_counter, cc_v1, krocc_v, rmse_v, stress1, dist1, y_true1, score_val = test( 201 | data_test_loader, net, loss) 202 | print('best performance: plcc{} srcc{}'.format(cc_v1, srocc_v1)) 203 | srocc_v2, total_val_loss, val_counter, cc_v2, krocc_v, rmse_v, stress2, dist2, y_true2, score_val = test( 204 | data_test_aligned_loader, net, loss) 205 | print('best performance in Pixel-wise aligned: plcc{} srcc{}'.format(cc_v2, srocc_v2)) 206 | srocc_v3, total_val_loss, val_counter, cc_v3, krocc_v, rmse_v, stress3, dist3, y_true3, score_val = test( 207 | data_test_notaligned_loader, net, loss) 208 | print('best performance in non-Pixel-wise aligned: plcc{} srcc{}'.format(cc_v3, srocc_v3)) 209 | return dist1, y_true1, stress1, cc_v1, srocc_v1, dist2, y_true2, stress2, cc_v2, srocc_v2, dist3, y_true3, stress3, cc_v3, srocc_v3 210 | 211 | 212 | def test(data_val_loader, net, loss): 213 | total_val_loss = 0 214 | val_counter = 0 215 | score_val = 0 216 | dist = [] 217 | y_true = [] 218 | device = torch.device("cuda") 219 | for i, data in enumerate(data_val_loader, 0): 220 | with torch.no_grad(): 221 | x, y, gts = data 222 | y_val = gts.numpy() 223 | x, y, gts = \ 224 | Variable(x).to(device), \ 225 | Variable(y).to(device), \ 226 | Variable(gts).to(device) 227 | 228 | score, _, _, _, _, _, _, _, _, _ = net(x, y) 229 | 230 | score_loss = loss(score, gts) 231 | loss_size = score_loss 232 | total_val_loss += loss_size.cpu().numpy() 233 | score_val = score_val + score_loss.item() 234 | val_counter += 1 235 | pred = (torch.squeeze(score)).cpu().detach().numpy().tolist() 236 | if isinstance(pred, list): 237 | dist.extend(pred) 238 | y_true.extend(y_val.tolist()) 239 | else: 240 | dist.append(np.array(pred)) 241 | y_true.append(y_val) 242 | # Calculate correlation coefficients between the predicted values and ground truth values on validation set. 243 | dist_np = np.array(dist).squeeze() 244 | y_true_np = np.array(y_true).squeeze() 245 | stress = compute_stress(dist_np, y_true_np) 246 | _, cc_v, srocc_v, krocc_v, rmse_v = coeff_fit(dist_np, y_true_np) 247 | return srocc_v, total_val_loss, val_counter, cc_v, krocc_v, rmse_v, stress, dist, y_true, score_val 248 | 249 | 250 | def calc_loss(log_p, logdet, image_size, n_bins): 251 | n_pixel = image_size * image_size * 3 252 | loss = -log(n_bins) * n_pixel 253 | loss = loss + logdet + log_p 254 | 255 | return (-loss / (log(2) * n_pixel)).mean(), (log_p / (log(2) * n_pixel)).mean(), ( 256 | logdet / (log(2) * n_pixel)).mean() 257 | -------------------------------------------------------------------------------- /flow.py: -------------------------------------------------------------------------------- 1 | import torch 2 | from torch import nn 3 | from torch.nn import functional as F 4 | from math import log, pi, exp 5 | import numpy as np 6 | from scipy import linalg as la 7 | logabs = lambda x: torch.log(torch.abs(x)) 8 | 9 | class ActNorm(nn.Module): 10 | def __init__(self, in_channel, logdet=True): 11 | super().__init__() 12 | 13 | self.loc = nn.Parameter(torch.zeros(1, in_channel, 1, 1)) 14 | self.scale = nn.Parameter(torch.ones(1, in_channel, 1, 1)) 15 | 16 | self.register_buffer("initialized", torch.tensor(0, dtype=torch.uint8)) 17 | self.logdet = logdet 18 | 19 | def initialize(self, input): 20 | with torch.no_grad(): 21 | flatten = input.permute(1, 0, 2, 3).contiguous().view(input.shape[1], -1) 22 | mean = ( 23 | flatten.mean(1) 24 | .unsqueeze(1) 25 | .unsqueeze(2) 26 | .unsqueeze(3) 27 | .permute(1, 0, 2, 3) 28 | ) 29 | std = ( 30 | flatten.std(1) 31 | .unsqueeze(1) 32 | .unsqueeze(2) 33 | .unsqueeze(3) 34 | .permute(1, 0, 2, 3) 35 | ) 36 | 37 | self.loc.data.copy_(-mean) 38 | self.scale.data.copy_(1 / (std + 1e-6)) 39 | 40 | def forward(self, input): 41 | _, _, height, width = input.shape 42 | 43 | if self.initialized.item() == 0: 44 | self.initialize(input) 45 | self.initialized.fill_(1) 46 | 47 | log_abs = logabs(self.scale) 48 | 49 | logdet = height * width * torch.sum(log_abs) 50 | 51 | if self.logdet: 52 | return self.scale * (input + self.loc), logdet 53 | 54 | else: 55 | return self.scale * (input + self.loc) 56 | 57 | def reverse(self, output): 58 | return output / self.scale - self.loc 59 | 60 | class InvConv2d(nn.Module): 61 | def __init__(self, in_channel): 62 | super().__init__() 63 | 64 | weight = torch.randn(in_channel, in_channel) 65 | q, _ = torch.qr(weight) 66 | weight = q.unsqueeze(2).unsqueeze(3) 67 | self.weight = nn.Parameter(weight) 68 | 69 | def forward(self, input): 70 | _, _, height, width = input.shape 71 | 72 | out = F.conv2d(input, self.weight) 73 | logdet = ( 74 | height * width * torch.slogdet(self.weight.squeeze().double())[1].float() 75 | ) 76 | 77 | return out, logdet 78 | 79 | def reverse(self, output): 80 | return F.conv2d( 81 | output, self.weight.squeeze().inverse().unsqueeze(2).unsqueeze(3) 82 | ) 83 | 84 | class InvConv2dLU(nn.Module): 85 | def __init__(self, in_channel): 86 | super().__init__() 87 | 88 | weight = np.random.randn(in_channel, in_channel) 89 | q, _ = la.qr(weight) 90 | w_p, w_l, w_u = la.lu(q.astype(np.float32)) 91 | w_s = np.diag(w_u) 92 | w_u = np.triu(w_u, 1) 93 | u_mask = np.triu(np.ones_like(w_u), 1) 94 | l_mask = u_mask.T 95 | 96 | w_p = torch.from_numpy(w_p) 97 | w_l = torch.from_numpy(w_l) 98 | w_s = torch.from_numpy(w_s) 99 | w_u = torch.from_numpy(w_u) 100 | 101 | self.register_buffer("w_p", w_p) 102 | self.register_buffer("u_mask", torch.from_numpy(u_mask)) 103 | self.register_buffer("l_mask", torch.from_numpy(l_mask)) 104 | self.register_buffer("s_sign", torch.sign(w_s)) 105 | self.register_buffer("l_eye", torch.eye(l_mask.shape[0])) 106 | self.w_l = nn.Parameter(w_l) 107 | self.w_s = nn.Parameter(logabs(w_s)) 108 | self.w_u = nn.Parameter(w_u) 109 | 110 | def forward(self, input): 111 | _, _, height, width = input.shape 112 | 113 | weight = self.calc_weight() 114 | 115 | out = F.conv2d(input, weight) 116 | logdet = height * width * torch.sum(self.w_s) 117 | 118 | return out, logdet 119 | 120 | def calc_weight(self): 121 | weight = ( 122 | self.w_p 123 | @ (self.w_l * self.l_mask + self.l_eye) 124 | @ ((self.w_u * self.u_mask) + torch.diag(self.s_sign * torch.exp(self.w_s))) 125 | ) 126 | 127 | return weight.unsqueeze(2).unsqueeze(3) 128 | 129 | def reverse(self, output): 130 | weight = self.calc_weight() 131 | 132 | return F.conv2d(output, weight.squeeze().inverse().unsqueeze(2).unsqueeze(3)) 133 | 134 | class ZeroConv2d(nn.Module): 135 | def __init__(self, in_channel, out_channel, padding=1): 136 | super().__init__() 137 | 138 | self.conv = nn.Conv2d(in_channel, out_channel, 3, padding=0) 139 | self.conv.weight.data.zero_() 140 | self.conv.bias.data.zero_() 141 | self.scale = nn.Parameter(torch.zeros(1, out_channel, 1, 1)) 142 | 143 | def forward(self, input): 144 | out = F.pad(input, [1, 1, 1, 1], value=1) 145 | out = self.conv(out) 146 | out = out * torch.exp(self.scale * 3) 147 | 148 | return out 149 | 150 | class AffineCoupling(nn.Module): 151 | def __init__(self, in_channel, filter_size=512, affine=True): 152 | super().__init__() 153 | 154 | self.affine = affine 155 | 156 | self.net = nn.Sequential( 157 | nn.Conv2d(in_channel // 2, filter_size, 3, padding=1), 158 | nn.ReLU(inplace=True), 159 | nn.Conv2d(filter_size, filter_size, 1), 160 | nn.ReLU(inplace=True), 161 | ZeroConv2d(filter_size, in_channel if self.affine else in_channel // 2), 162 | ) 163 | 164 | self.net[0].weight.data.normal_(0, 0.05) 165 | self.net[0].bias.data.zero_() 166 | 167 | self.net[2].weight.data.normal_(0, 0.05) 168 | self.net[2].bias.data.zero_() 169 | 170 | def forward(self, input): 171 | in_a, in_b = input.chunk(2, 1) 172 | 173 | if self.affine: 174 | log_s, t = self.net(in_a).chunk(2, 1) 175 | s = F.sigmoid(log_s + 2) 176 | out_b = (in_b + t) * s 177 | 178 | logdet = torch.sum(torch.log(s).view(input.shape[0], -1), 1) 179 | 180 | else: 181 | net_out = self.net(in_a) 182 | out_b = in_b + net_out 183 | logdet = None 184 | 185 | return torch.cat([in_a, out_b], 1), logdet 186 | 187 | def reverse(self, output): 188 | out_a, out_b = output.chunk(2, 1) 189 | 190 | if self.affine: 191 | log_s, t = self.net(out_a).chunk(2, 1) 192 | # s = torch.exp(log_s) 193 | s = F.sigmoid(log_s + 2) 194 | # in_a = (out_a - t) / s 195 | in_b = out_b / s - t 196 | 197 | else: 198 | net_out = self.net(out_a) 199 | in_b = out_b - net_out 200 | 201 | return torch.cat([out_a, in_b], 1) 202 | 203 | class Flow(nn.Module): 204 | def __init__(self, in_channel, affine=True, conv_lu=True): 205 | super().__init__() 206 | 207 | self.actnorm = ActNorm(in_channel) 208 | 209 | if conv_lu: 210 | self.invconv = InvConv2dLU(in_channel) 211 | 212 | else: 213 | self.invconv = InvConv2d(in_channel) 214 | 215 | self.coupling = AffineCoupling(in_channel, affine=affine) 216 | 217 | def forward(self, input): 218 | out, logdet = self.actnorm(input) 219 | out, det1 = self.invconv(out) 220 | out, det2 = self.coupling(out) 221 | 222 | logdet = logdet + det1 223 | if det2 is not None: 224 | logdet = logdet + det2 225 | 226 | return out, logdet 227 | 228 | def reverse(self, output): 229 | input = self.coupling.reverse(output) 230 | input = self.invconv.reverse(input) 231 | input = self.actnorm.reverse(input) 232 | 233 | return input 234 | 235 | def gaussian_log_p(x, mean, log_sd): 236 | return -0.5 * log(2 * pi) - log_sd - 0.5 * (x - mean) ** 2 / torch.exp(2 * log_sd) 237 | 238 | def gaussian_sample(eps, mean, log_sd): 239 | return mean + torch.exp(log_sd) * eps 240 | 241 | class Block(nn.Module): 242 | def __init__(self, in_channel, n_flow, split=True, affine=True, conv_lu=True): 243 | super().__init__() 244 | 245 | squeeze_dim = in_channel * 4 246 | 247 | self.flows = nn.ModuleList() 248 | for i in range(n_flow): 249 | self.flows.append(Flow(squeeze_dim, affine=affine, conv_lu=conv_lu)) 250 | 251 | self.split = split 252 | 253 | if split: 254 | self.prior = ZeroConv2d(in_channel * 2, in_channel * 4) 255 | 256 | else: 257 | self.prior = ZeroConv2d(in_channel * 4, in_channel * 8) 258 | 259 | def forward(self, input): 260 | b_size, n_channel, height, width = input.shape 261 | squeezed = input.view(b_size, n_channel, height // 2, 2, width // 2, 2) 262 | squeezed = squeezed.permute(0, 1, 3, 5, 2, 4) 263 | out = squeezed.contiguous().view(b_size, n_channel * 4, height // 2, width // 2) 264 | 265 | logdet = 0 266 | 267 | for flow in self.flows: 268 | out, det = flow(out) 269 | logdet = logdet + det 270 | 271 | 272 | if self.split: 273 | out, z_new = out.chunk(2, 1) 274 | mean, log_sd = self.prior(out).chunk(2, 1) 275 | 276 | log_p = gaussian_log_p(z_new, mean, log_sd) 277 | log_p = log_p.view(b_size, -1).sum(1) 278 | 279 | else: 280 | one = torch.ones_like(out) 281 | 282 | mean, log_sd = self.prior(one).chunk(2, 1) 283 | 284 | log_p = gaussian_log_p(out, mean, log_sd) 285 | log_p = log_p.view(b_size, -1).sum(1) 286 | z_new = out 287 | 288 | #self.log_sd = log_sd 289 | 290 | return out, logdet, log_p, z_new 291 | 292 | def reverse(self, output, eps=None, reconstruct=False): 293 | input = output 294 | 295 | if reconstruct: 296 | if self.split: 297 | input = torch.cat([output, eps], 1) ## channel-wise concat 298 | 299 | else: 300 | input = eps 301 | 302 | else: 303 | if self.split: 304 | mean, log_sd = self.prior(input).chunk(2, 1) 305 | z = gaussian_sample(eps, mean, log_sd) 306 | input = torch.cat([output, z], 1) 307 | 308 | else: 309 | one = torch.ones_like(input) 310 | mean, log_sd = self.prior(one).chunk(2, 1) 311 | z = gaussian_sample(eps, mean, log_sd) 312 | input = z 313 | 314 | for flow in self.flows[::-1]: 315 | input = flow.reverse(input) 316 | 317 | b_size, n_channel, height, width = input.shape 318 | 319 | unsqueezed = input.view(b_size, n_channel // 4, 2, 2, height, width) 320 | unsqueezed = unsqueezed.permute(0, 1, 4, 2, 5, 3) 321 | unsqueezed = unsqueezed.contiguous().view( 322 | b_size, n_channel // 4, height * 2, width * 2 323 | ) 324 | 325 | return unsqueezed 326 | 327 | class Glow(nn.Module): 328 | def __init__( 329 | self, in_channel, n_flow, n_block, affine=True, conv_lu=True 330 | ): 331 | super().__init__() 332 | 333 | self.blocks = nn.ModuleList() 334 | n_channel = in_channel 335 | for i in range(n_block - 1): 336 | self.blocks.append(Block(n_channel, n_flow, affine=affine, conv_lu=conv_lu)) 337 | n_channel *= 2 338 | self.blocks.append(Block(n_channel, n_flow, split=False, affine=affine)) 339 | 340 | def forward(self, input): 341 | log_p_sum = 0 342 | logdet = 0 343 | out = input 344 | z_outs = [] 345 | 346 | for i, block in enumerate(self.blocks): 347 | out, det, log_p, z_new = block(out) 348 | z_outs.append(z_new) 349 | logdet = logdet + det 350 | 351 | if log_p is not None: 352 | log_p_sum = log_p_sum + log_p 353 | 354 | return log_p_sum, logdet, z_outs 355 | 356 | def reverse(self, z_list, reconstruct=True, cd_map=False): 357 | for i, block in enumerate(self.blocks[::-1]): 358 | if i == 0: 359 | input = block.reverse(z_list[-1], z_list[-1], reconstruct=reconstruct) 360 | 361 | else: 362 | input = block.reverse(input, z_list[-(i + 1)], reconstruct=reconstruct) 363 | 364 | return input 365 | -------------------------------------------------------------------------------- /data/1/test_notaligned.csv: -------------------------------------------------------------------------------- 1 | I0006_01_03.png,I0006_01_04.png,4.585466157 2 | I0006_01_02.png,I0006_01_06.png,3.518305565 3 | I0006_01_02.png,I0006_01_03.png,3.621615939 4 | I0006_01_01.png,I0006_01_02.png,3.480775818 5 | I0006_01_01.png,I0006_01_03.png,3.271868452 6 | I0006_01_01.png,I0006_01_04.png,4.57595483 7 | I0006_01_04.png,I0006_01_05.png,3.070782682 8 | I0006_01_02.png,I0006_01_05.png,5.72438561 9 | I0006_01_03.png,I0006_01_06.png,5.9147982 10 | I0006_01_01.png,I0006_01_05.png,5.287737329 11 | I0006_01_04.png,I0006_01_06.png,6.295935194 12 | I0006_01_02.png,I0006_01_04.png,6.097630454 13 | I0006_01_03.png,I0006_01_05.png,2.462734718 14 | I0006_01_01.png,I0006_01_06.png,3.866461458 15 | I0006_01_05.png,I0006_01_06.png,8.666758489 16 | I0013_01_03.png,I0013_01_05.png,4.070132905 17 | I0013_01_03.png,I0013_01_04.png,4.174690853 18 | I0013_01_02.png,I0013_01_03.png,3.24520533 19 | I0013_01_03.png,I0013_01_06.png,8.724584699 20 | I0013_01_04.png,I0013_01_05.png,6.882454695 21 | I0013_01_01.png,I0013_01_04.png,2.437602498 22 | I0013_01_02.png,I0013_01_05.png,3.199226198 23 | I0013_01_01.png,I0013_01_03.png,4.084634631 24 | I0013_01_04.png,I0013_01_06.png,10.09526255 25 | I0013_01_02.png,I0013_01_04.png,4.128733237 26 | I0013_01_01.png,I0013_01_05.png,6.137289293 27 | I0013_01_01.png,I0013_01_06.png,9.568328754 28 | I0013_01_02.png,I0013_01_06.png,5.291757213 29 | I0013_01_01.png,I0013_01_02.png,4.487085133 30 | I0013_01_05.png,I0013_01_06.png,3.78551412 31 | I0027_01_04.png,I0027_01_06.png,13.57734229 32 | I0027_01_04.png,I0027_01_05.png,14.73093229 33 | I0027_01_05.png,I0027_01_06.png,2.819281637 34 | I0027_01_02.png,I0027_01_05.png,5.876400828 35 | I0027_01_03.png,I0027_01_04.png,14.78235272 36 | I0027_01_03.png,I0027_01_05.png,3.814945273 37 | I0027_01_02.png,I0027_01_03.png,8.031100296 38 | I0027_01_01.png,I0027_01_03.png,6.08837392 39 | I0027_01_02.png,I0027_01_04.png,11.66892938 40 | I0027_01_01.png,I0027_01_02.png,2.843821905 41 | I0027_01_01.png,I0027_01_04.png,12.87468692 42 | I0027_01_02.png,I0027_01_06.png,4.571191661 43 | I0027_01_01.png,I0027_01_05.png,3.565609564 44 | I0027_01_01.png,I0027_01_06.png,2.937161431 45 | I0027_01_03.png,I0027_01_06.png,5.210823865 46 | I0038_01_01.png,I0038_01_03.png,5.024905758 47 | I0038_01_04.png,I0038_01_05.png,6.057085341 48 | I0038_01_01.png,I0038_01_02.png,6.101407987 49 | I0038_01_02.png,I0038_01_03.png,6.367617247 50 | I0038_01_03.png,I0038_01_04.png,6.161388306 51 | I0038_01_05.png,I0038_01_06.png,6.128229257 52 | I0038_01_01.png,I0038_01_05.png,8.381604506 53 | I0038_01_03.png,I0038_01_06.png,10.93634089 54 | I0038_01_02.png,I0038_01_06.png,6.89159407 55 | I0038_01_03.png,I0038_01_05.png,8.280428422 56 | I0038_01_01.png,I0038_01_06.png,12.45383736 57 | I0038_01_02.png,I0038_01_04.png,3.999168898 58 | I0038_01_01.png,I0038_01_04.png,8.526542921 59 | I0038_01_02.png,I0038_01_05.png,4.034156275 60 | I0038_01_04.png,I0038_01_06.png,6.271187115 61 | I0040_01_03.png,I0040_01_06.png,3.248361659 62 | I0040_01_01.png,I0040_01_02.png,5.230226228 63 | I0040_01_01.png,I0040_01_03.png,4.675074368 64 | I0040_01_01.png,I0040_01_05.png,4.220208518 65 | I0040_01_04.png,I0040_01_06.png,2.326859353 66 | I0040_01_03.png,I0040_01_04.png,3.341948818 67 | I0040_01_05.png,I0040_01_06.png,3.220545109 68 | I0040_01_02.png,I0040_01_03.png,4.3912304 69 | I0040_01_02.png,I0040_01_05.png,3.322786291 70 | I0040_01_01.png,I0040_01_04.png,4.726180761 71 | I0040_01_02.png,I0040_01_06.png,2.730122565 72 | I0040_01_04.png,I0040_01_05.png,3.110429419 73 | I0040_01_01.png,I0040_01_06.png,4.328199053 74 | I0040_01_02.png,I0040_01_04.png,3.548821795 75 | I0040_01_03.png,I0040_01_05.png,3.275827452 76 | I0045_01_02.png,I0045_01_04.png,2.522010415 77 | I0045_01_01.png,I0045_01_02.png,2.015738851 78 | I0045_01_02.png,I0045_01_03.png,2.786622418 79 | I0045_01_04.png,I0045_01_06.png,2.040580476 80 | I0045_01_04.png,I0045_01_05.png,2.842808278 81 | I0045_01_01.png,I0045_01_03.png,2.756349347 82 | I0045_01_01.png,I0045_01_04.png,1.868231683 83 | I0045_01_03.png,I0045_01_06.png,2.107080777 84 | I0045_01_03.png,I0045_01_04.png,2.768369568 85 | I0045_01_01.png,I0045_01_05.png,2.775708006 86 | I0045_01_05.png,I0045_01_06.png,2.937761998 87 | I0045_01_02.png,I0045_01_06.png,2.738557889 88 | I0045_01_02.png,I0045_01_05.png,1.984175092 89 | I0045_01_01.png,I0045_01_06.png,2.717623036 90 | I0045_01_03.png,I0045_01_05.png,3.1581156 91 | I0049_01_05.png,I0049_01_06.png,4.533052966 92 | I0049_01_02.png,I0049_01_06.png,3.340797513 93 | I0049_01_04.png,I0049_01_05.png,4.457137285 94 | I0049_01_03.png,I0049_01_04.png,3.102678993 95 | I0049_01_02.png,I0049_01_04.png,1.670506468 96 | I0049_01_02.png,I0049_01_03.png,2.927506292 97 | I0049_01_01.png,I0049_01_06.png,2.494857315 98 | I0049_01_01.png,I0049_01_02.png,3.651272329 99 | I0049_01_03.png,I0049_01_05.png,3.907928305 100 | I0049_01_01.png,I0049_01_05.png,3.396163461 101 | I0049_01_01.png,I0049_01_04.png,3.115334367 102 | I0049_01_02.png,I0049_01_05.png,4.182854041 103 | I0049_01_04.png,I0049_01_06.png,4.215781264 104 | I0049_01_01.png,I0049_01_03.png,3.024191335 105 | I0049_01_03.png,I0049_01_06.png,2.911618494 106 | I0050_01_04.png,I0050_01_05.png,6.042764371 107 | I0050_01_01.png,I0050_01_05.png,5.612320002 108 | I0050_01_01.png,I0050_01_04.png,3.298840801 109 | I0050_01_01.png,I0050_01_02.png,3.424331102 110 | I0050_01_04.png,I0050_01_06.png,5.081228051 111 | I0050_01_03.png,I0050_01_06.png,7.822358851 112 | I0050_01_02.png,I0050_01_06.png,4.122502484 113 | I0050_01_01.png,I0050_01_06.png,4.740840435 114 | I0050_01_03.png,I0050_01_04.png,4.039823893 115 | I0050_01_03.png,I0050_01_05.png,8.452700889 116 | I0050_01_02.png,I0050_01_05.png,4.283280396 117 | I0050_01_02.png,I0050_01_03.png,6.238668221 118 | I0050_01_05.png,I0050_01_06.png,3.722552737 119 | I0050_01_02.png,I0050_01_04.png,3.591375041 120 | I0050_01_01.png,I0050_01_03.png,4.42735239 121 | I0051_01_03.png,I0051_01_05.png,3.92281263 122 | I0051_01_01.png,I0051_01_05.png,7.251805018 123 | I0051_01_01.png,I0051_01_04.png,5.183663665 124 | I0051_01_02.png,I0051_01_06.png,2.976825293 125 | I0051_01_02.png,I0051_01_04.png,4.503028863 126 | I0051_01_04.png,I0051_01_05.png,2.972287633 127 | I0051_01_01.png,I0051_01_06.png,3.394369735 128 | I0051_01_04.png,I0051_01_06.png,3.364908692 129 | I0051_01_02.png,I0051_01_03.png,5.069286208 130 | I0051_01_05.png,I0051_01_06.png,5.016534816 131 | I0051_01_03.png,I0051_01_04.png,3.290957375 132 | I0051_01_02.png,I0051_01_05.png,5.515267846 133 | I0051_01_03.png,I0051_01_06.png,5.500025805 134 | I0051_01_01.png,I0051_01_02.png,4.252508104 135 | I0051_01_01.png,I0051_01_03.png,3.933072734 136 | I0052_01_03.png,I0052_01_04.png,3.00875157 137 | I0052_01_04.png,I0052_01_06.png,7.118981725 138 | I0052_01_03.png,I0052_01_06.png,6.820963607 139 | I0052_01_01.png,I0052_01_05.png,6.287509345 140 | I0052_01_02.png,I0052_01_05.png,3.437789857 141 | I0052_01_01.png,I0052_01_04.png,7.329454675 142 | I0052_01_03.png,I0052_01_05.png,3.103628351 143 | I0052_01_01.png,I0052_01_02.png,5.788106591 144 | I0052_01_01.png,I0052_01_06.png,10.2639703 145 | I0052_01_05.png,I0052_01_06.png,5.28505352 146 | I0052_01_02.png,I0052_01_06.png,5.517380688 147 | I0052_01_02.png,I0052_01_04.png,2.907932707 148 | I0052_01_04.png,I0052_01_05.png,2.911475577 149 | I0052_01_02.png,I0052_01_03.png,3.216443022 150 | I0052_01_01.png,I0052_01_03.png,5.053518352 151 | I0054_01_01.png,I0054_01_04.png,4.317497126 152 | I0054_01_04.png,I0054_01_05.png,4.668220048 153 | I0054_01_01.png,I0054_01_03.png,2.476638666 154 | I0054_01_02.png,I0054_01_03.png,5.351037368 155 | I0054_01_02.png,I0054_01_06.png,9.48944602 156 | I0054_01_03.png,I0054_01_06.png,7.025255132 157 | I0054_01_05.png,I0054_01_06.png,7.011200739 158 | I0054_01_04.png,I0054_01_06.png,8.23164926 159 | I0054_01_01.png,I0054_01_02.png,6.286698534 160 | I0054_01_02.png,I0054_01_04.png,2.011586736 161 | I0054_01_01.png,I0054_01_05.png,3.412180691 162 | I0054_01_01.png,I0054_01_06.png,7.694264434 163 | I0054_01_03.png,I0054_01_05.png,2.95041264 164 | I0054_01_03.png,I0054_01_04.png,3.797154293 165 | I0054_01_02.png,I0054_01_05.png,5.335018474 166 | I0056_01_02.png,I0056_01_04.png,7.432108594 167 | I0056_01_02.png,I0056_01_05.png,3.537351111 168 | I0056_01_01.png,I0056_01_02.png,5.437067279 169 | I0056_01_01.png,I0056_01_04.png,6.59910507 170 | I0056_01_03.png,I0056_01_05.png,3.466493294 171 | I0056_01_02.png,I0056_01_06.png,6.345681204 172 | I0056_01_02.png,I0056_01_03.png,6.729641506 173 | I0056_01_04.png,I0056_01_06.png,3.086837524 174 | I0056_01_01.png,I0056_01_05.png,2.244065532 175 | I0056_01_01.png,I0056_01_06.png,6.176208152 176 | I0056_01_05.png,I0056_01_06.png,4.212373677 177 | I0056_01_03.png,I0056_01_06.png,2.675504043 178 | I0056_01_03.png,I0056_01_04.png,3.297941554 179 | I0056_01_01.png,I0056_01_03.png,4.9984739 180 | I0056_01_04.png,I0056_01_05.png,6.230727429 181 | I0060_01_01.png,I0060_01_04.png,6.060247382 182 | I0060_01_03.png,I0060_01_04.png,2.383369962 183 | I0060_01_01.png,I0060_01_06.png,4.32204365 184 | I0060_01_02.png,I0060_01_04.png,6.037965384 185 | I0060_01_01.png,I0060_01_02.png,3.402996034 186 | I0060_01_05.png,I0060_01_06.png,4.523647132 187 | I0060_01_03.png,I0060_01_05.png,6.552737597 188 | I0060_01_02.png,I0060_01_06.png,2.927734542 189 | I0060_01_01.png,I0060_01_03.png,5.838771425 190 | I0060_01_02.png,I0060_01_05.png,4.799887284 191 | I0060_01_04.png,I0060_01_06.png,3.125123565 192 | I0060_01_04.png,I0060_01_05.png,7.287144814 193 | I0060_01_03.png,I0060_01_06.png,4.019773347 194 | I0060_01_01.png,I0060_01_05.png,5.560047815 195 | I0060_01_02.png,I0060_01_03.png,6.661547257 196 | I0064_01_03.png,I0064_01_06.png,2.698033169 197 | I0064_01_02.png,I0064_01_03.png,3.800147978 198 | I0064_01_02.png,I0064_01_06.png,3.366760241 199 | I0064_01_04.png,I0064_01_06.png,6.203763987 200 | I0064_01_01.png,I0064_01_04.png,6.413359487 201 | I0064_01_03.png,I0064_01_04.png,5.805196441 202 | I0064_01_05.png,I0064_01_06.png,2.146792465 203 | I0064_01_01.png,I0064_01_05.png,3.460110833 204 | I0064_01_02.png,I0064_01_04.png,3.98363469 205 | I0064_01_01.png,I0064_01_02.png,4.085589736 206 | I0064_01_02.png,I0064_01_05.png,2.927789275 207 | I0064_01_04.png,I0064_01_05.png,7.005813196 208 | I0064_01_03.png,I0064_01_05.png,2.327566826 209 | I0064_01_01.png,I0064_01_03.png,3.315970693 210 | I0064_01_01.png,I0064_01_06.png,3.083391265 211 | I0069_01_01.png,I0069_01_02.png,2.947091516 212 | I0069_01_04.png,I0069_01_06.png,2.417783971 213 | I0069_01_01.png,I0069_01_04.png,3.60699706 214 | I0069_01_01.png,I0069_01_03.png,2.333273714 215 | I0069_01_01.png,I0069_01_05.png,2.370963245 216 | I0069_01_04.png,I0069_01_05.png,3.76279305 217 | I0069_01_02.png,I0069_01_03.png,2.670028713 218 | I0069_01_01.png,I0069_01_06.png,2.60314255 219 | I0069_01_02.png,I0069_01_05.png,2.873338766 220 | I0069_01_02.png,I0069_01_06.png,3.497619519 221 | I0069_01_02.png,I0069_01_04.png,4.425125464 222 | I0069_01_03.png,I0069_01_05.png,1.738514475 223 | I0069_01_03.png,I0069_01_06.png,2.049444247 224 | I0069_01_03.png,I0069_01_04.png,3.235753333 225 | I0069_01_05.png,I0069_01_06.png,1.828914524 226 | I0083_01_01.png,I0083_01_04.png,4.847960391 227 | I0083_01_02.png,I0083_01_05.png,2.097618458 228 | I0083_01_01.png,I0083_01_03.png,2.913878968 229 | I0083_01_03.png,I0083_01_05.png,2.009360428 230 | I0083_01_01.png,I0083_01_06.png,2.500326607 231 | I0083_01_02.png,I0083_01_03.png,3.532573999 232 | I0083_01_01.png,I0083_01_02.png,1.797701088 233 | I0083_01_03.png,I0083_01_06.png,2.647883796 234 | I0083_01_02.png,I0083_01_04.png,3.888447649 235 | I0083_01_05.png,I0083_01_06.png,2.279819816 236 | I0083_01_03.png,I0083_01_04.png,3.32015243 237 | I0083_01_04.png,I0083_01_06.png,2.706624342 238 | I0083_01_01.png,I0083_01_05.png,2.580571584 239 | I0083_01_02.png,I0083_01_06.png,2.195963407 240 | I0083_01_04.png,I0083_01_05.png,2.804007181 241 | I0093_01_02.png,I0093_01_06.png,4.738798348 242 | I0093_01_01.png,I0093_01_05.png,2.959776515 243 | I0093_01_04.png,I0093_01_06.png,4.291638018 244 | I0093_01_01.png,I0093_01_03.png,3.882759164 245 | I0093_01_01.png,I0093_01_02.png,2.487950962 246 | I0093_01_05.png,I0093_01_06.png,3.808936766 247 | I0093_01_03.png,I0093_01_05.png,2.8727664 248 | I0093_01_04.png,I0093_01_05.png,3.470687268 249 | I0093_01_02.png,I0093_01_03.png,3.30747067 250 | I0093_01_02.png,I0093_01_05.png,3.583537771 251 | I0093_01_02.png,I0093_01_04.png,2.745767452 252 | I0093_01_01.png,I0093_01_04.png,2.659427154 253 | I0093_01_01.png,I0093_01_06.png,4.777315545 254 | I0093_01_03.png,I0093_01_04.png,2.545979086 255 | I0093_01_03.png,I0093_01_06.png,3.393077558 256 | I0097_01_05.png,I0097_01_06.png,10.29768371 257 | I0097_01_02.png,I0097_01_04.png,11.17725659 258 | I0097_01_04.png,I0097_01_06.png,12.4943706 259 | I0097_01_03.png,I0097_01_04.png,7.059294394 260 | I0097_01_02.png,I0097_01_06.png,3.005343715 261 | I0097_01_01.png,I0097_01_04.png,3.884966522 262 | I0097_01_01.png,I0097_01_03.png,4.664951289 263 | I0097_01_03.png,I0097_01_05.png,4.212715041 264 | I0097_01_01.png,I0097_01_06.png,12.36189565 265 | I0097_01_03.png,I0097_01_06.png,11.28768467 266 | I0097_01_02.png,I0097_01_03.png,9.808164061 267 | I0097_01_01.png,I0097_01_05.png,4.048001547 268 | I0097_01_01.png,I0097_01_02.png,9.080574604 269 | I0097_01_04.png,I0097_01_05.png,5.118670053 270 | I0097_01_02.png,I0097_01_05.png,8.632639799 271 | I0104_01_01.png,I0104_01_06.png,9.499349516 272 | I0104_01_01.png,I0104_01_04.png,5.257884188 273 | I0104_01_01.png,I0104_01_05.png,5.73983834 274 | I0104_01_02.png,I0104_01_04.png,5.348587593 275 | I0104_01_02.png,I0104_01_03.png,3.100376748 276 | I0104_01_01.png,I0104_01_03.png,3.686525553 277 | I0104_01_02.png,I0104_01_05.png,5.398826995 278 | I0104_01_03.png,I0104_01_05.png,4.281784626 279 | I0104_01_01.png,I0104_01_02.png,3.464887766 280 | I0104_01_05.png,I0104_01_06.png,5.750458485 281 | I0104_01_03.png,I0104_01_04.png,5.743424603 282 | I0104_01_04.png,I0104_01_05.png,7.858555797 283 | I0104_01_03.png,I0104_01_06.png,7.782654033 284 | I0104_01_02.png,I0104_01_06.png,9.121867466 285 | I0104_01_04.png,I0104_01_06.png,11.27789162 286 | I0107_01_02.png,I0107_01_03.png,7.122779779 287 | I0107_01_03.png,I0107_01_05.png,3.623435956 288 | I0107_01_02.png,I0107_01_04.png,3.423806431 289 | I0107_01_01.png,I0107_01_05.png,3.079806754 290 | I0107_01_03.png,I0107_01_04.png,8.456887141 291 | I0107_01_02.png,I0107_01_05.png,4.995876921 292 | I0107_01_03.png,I0107_01_06.png,3.938673891 293 | I0107_01_04.png,I0107_01_05.png,6.905097943 294 | I0107_01_02.png,I0107_01_06.png,9.626228936 295 | I0107_01_01.png,I0107_01_04.png,6.089045161 296 | I0107_01_04.png,I0107_01_06.png,10.24948786 297 | I0107_01_05.png,I0107_01_06.png,5.714682288 298 | I0107_01_01.png,I0107_01_03.png,4.917691053 299 | I0107_01_01.png,I0107_01_02.png,4.673155229 300 | I0107_01_01.png,I0107_01_06.png,7.759980168 301 | I0113_01_01.png,I0113_01_05.png,2.022060725 302 | I0113_01_02.png,I0113_01_06.png,2.733951893 303 | I0113_01_02.png,I0113_01_04.png,3.442706135 304 | I0113_01_04.png,I0113_01_05.png,2.994907925 305 | I0113_01_01.png,I0113_01_04.png,2.215669597 306 | I0113_01_05.png,I0113_01_06.png,4.175876532 307 | I0113_01_01.png,I0113_01_06.png,4.440644731 308 | I0113_01_03.png,I0113_01_06.png,2.818419986 309 | I0113_01_02.png,I0113_01_03.png,2.400411478 310 | I0113_01_03.png,I0113_01_05.png,3.082429194 311 | I0113_01_03.png,I0113_01_04.png,3.367581235 312 | I0113_01_01.png,I0113_01_03.png,3.037635527 313 | I0113_01_02.png,I0113_01_05.png,4.177590457 314 | I0113_01_01.png,I0113_01_02.png,3.680292239 315 | I0113_01_04.png,I0113_01_06.png,5.261292083 316 | I0117_01_02.png,I0117_01_05.png,2.713686089 317 | I0117_01_01.png,I0117_01_05.png,4.050357542 318 | I0117_01_01.png,I0117_01_02.png,3.256298694 319 | I0117_01_03.png,I0117_01_04.png,2.837951373 320 | I0117_01_02.png,I0117_01_04.png,3.35240282 321 | I0117_01_03.png,I0117_01_05.png,3.602677123 322 | I0117_01_02.png,I0117_01_03.png,3.228472802 323 | I0117_01_01.png,I0117_01_06.png,3.361790954 324 | I0117_01_04.png,I0117_01_05.png,3.492409612 325 | I0117_01_04.png,I0117_01_06.png,3.062647472 326 | I0117_01_03.png,I0117_01_06.png,3.190822119 327 | I0117_01_01.png,I0117_01_04.png,2.718577658 328 | I0117_01_01.png,I0117_01_03.png,3.359477402 329 | I0117_01_05.png,I0117_01_06.png,2.328343366 330 | I0117_01_02.png,I0117_01_06.png,2.334417519 331 | I0122_01_03.png,I0122_01_06.png,5.916806693 332 | I0122_01_01.png,I0122_01_06.png,7.491872227 333 | I0122_01_02.png,I0122_01_05.png,7.132449817 334 | I0122_01_01.png,I0122_01_03.png,5.078462044 335 | I0122_01_03.png,I0122_01_05.png,5.628771145 336 | I0122_01_01.png,I0122_01_05.png,5.309207046 337 | I0122_01_01.png,I0122_01_04.png,6.794312952 338 | I0122_01_05.png,I0122_01_06.png,5.94928839 339 | I0122_01_01.png,I0122_01_02.png,3.887567992 340 | I0122_01_02.png,I0122_01_03.png,5.278567079 341 | I0122_01_03.png,I0122_01_04.png,6.624859676 342 | I0122_01_02.png,I0122_01_06.png,7.520865694 343 | I0122_01_04.png,I0122_01_05.png,7.576407653 344 | I0122_01_04.png,I0122_01_06.png,8.938203447 345 | I0122_01_02.png,I0122_01_04.png,3.554191735 346 | I0124_01_01.png,I0124_01_05.png,6.336281213 347 | I0124_01_01.png,I0124_01_06.png,4.270999617 348 | I0124_01_03.png,I0124_01_04.png,2.698351415 349 | I0124_01_02.png,I0124_01_03.png,3.893446176 350 | I0124_01_02.png,I0124_01_05.png,6.951853933 351 | I0124_01_03.png,I0124_01_05.png,8.856749452 352 | I0124_01_01.png,I0124_01_04.png,5.167569351 353 | I0124_01_02.png,I0124_01_06.png,4.526609119 354 | I0124_01_01.png,I0124_01_02.png,2.258845869 355 | I0124_01_01.png,I0124_01_03.png,4.625805711 356 | I0124_01_04.png,I0124_01_05.png,7.922801634 357 | I0124_01_03.png,I0124_01_06.png,5.357573422 358 | I0124_01_04.png,I0124_01_06.png,6.960912032 359 | I0124_01_02.png,I0124_01_04.png,4.304133955 360 | I0124_01_05.png,I0124_01_06.png,5.721790771 361 | I0125_01_01.png,I0125_01_06.png,8.709569551 362 | I0125_01_02.png,I0125_01_03.png,5.171388805 363 | I0125_01_03.png,I0125_01_06.png,11.99065558 364 | I0125_01_02.png,I0125_01_04.png,5.00986137 365 | I0125_01_02.png,I0125_01_05.png,4.352927163 366 | I0125_01_01.png,I0125_01_04.png,3.774497604 367 | I0125_01_04.png,I0125_01_05.png,4.349741678 368 | I0125_01_05.png,I0125_01_06.png,9.124331955 369 | I0125_01_01.png,I0125_01_05.png,3.061463314 370 | I0125_01_03.png,I0125_01_05.png,3.649988908 371 | I0125_01_04.png,I0125_01_06.png,13.21802375 372 | I0125_01_01.png,I0125_01_02.png,3.719720605 373 | I0125_01_03.png,I0125_01_04.png,3.309163452 374 | I0125_01_01.png,I0125_01_03.png,3.447940733 375 | I0125_01_02.png,I0125_01_06.png,9.899572218 376 | I0126_01_03.png,I0126_01_05.png,3.166733027 377 | I0126_01_04.png,I0126_01_05.png,4.468304652 378 | I0126_01_03.png,I0126_01_04.png,4.342170998 379 | I0126_01_02.png,I0126_01_05.png,12.4400554 380 | I0126_01_01.png,I0126_01_05.png,2.349513645 381 | I0126_01_01.png,I0126_01_04.png,4.302671397 382 | I0126_01_01.png,I0126_01_03.png,3.589858795 383 | I0126_01_04.png,I0126_01_06.png,8.847938098 384 | I0126_01_02.png,I0126_01_03.png,14.41873181 385 | I0126_01_03.png,I0126_01_06.png,10.98053695 386 | I0126_01_01.png,I0126_01_02.png,13.78382453 387 | I0126_01_02.png,I0126_01_04.png,13.64186087 388 | I0126_01_05.png,I0126_01_06.png,8.75611856 389 | I0126_01_02.png,I0126_01_06.png,5.606730509 390 | I0126_01_01.png,I0126_01_06.png,7.339746057 391 | I0128_01_04.png,I0128_01_06.png,6.130803209 392 | I0128_01_04.png,I0128_01_05.png,7.44801763 393 | I0128_01_01.png,I0128_01_02.png,3.791462581 394 | I0128_01_02.png,I0128_01_03.png,3.826276327 395 | I0128_01_03.png,I0128_01_04.png,3.327476344 396 | I0128_01_02.png,I0128_01_06.png,5.255605668 397 | I0128_01_01.png,I0128_01_05.png,3.125058226 398 | I0128_01_01.png,I0128_01_04.png,5.073329258 399 | I0128_01_01.png,I0128_01_06.png,5.270293639 400 | I0128_01_02.png,I0128_01_04.png,3.715272738 401 | I0128_01_02.png,I0128_01_05.png,5.872025122 402 | I0128_01_05.png,I0128_01_06.png,5.614783964 403 | I0128_01_03.png,I0128_01_06.png,4.918550343 404 | I0128_01_01.png,I0128_01_03.png,4.481522236 405 | I0128_01_03.png,I0128_01_05.png,6.696248311 406 | I0130_01_02.png,I0130_01_04.png,2.412905856 407 | I0130_01_05.png,I0130_01_06.png,5.14741854 408 | I0130_01_02.png,I0130_01_05.png,4.439883573 409 | I0130_01_01.png,I0130_01_06.png,5.793651244 410 | I0130_01_01.png,I0130_01_05.png,4.305454259 411 | I0130_01_03.png,I0130_01_04.png,4.771139456 412 | I0130_01_02.png,I0130_01_06.png,7.507596538 413 | I0130_01_02.png,I0130_01_03.png,4.778581944 414 | I0130_01_01.png,I0130_01_04.png,3.771088145 415 | I0130_01_01.png,I0130_01_03.png,4.350566194 416 | I0130_01_03.png,I0130_01_06.png,4.40793368 417 | I0130_01_03.png,I0130_01_05.png,5.767010629 418 | I0130_01_04.png,I0130_01_06.png,8.496035498 419 | I0130_01_04.png,I0130_01_05.png,5.827192508 420 | I0130_01_01.png,I0130_01_02.png,3.320851097 421 | I0144_01_03.png,I0144_01_06.png,6.91673052 422 | I0144_01_01.png,I0144_01_05.png,4.785119643 423 | I0144_01_03.png,I0144_01_04.png,4.126524506 424 | I0144_01_01.png,I0144_01_06.png,8.204180474 425 | I0144_01_04.png,I0144_01_06.png,8.731417467 426 | I0144_01_03.png,I0144_01_05.png,5.00236711 427 | I0144_01_01.png,I0144_01_03.png,3.654102889 428 | I0144_01_04.png,I0144_01_05.png,4.863166284 429 | I0144_01_01.png,I0144_01_04.png,3.03730358 430 | I0144_01_01.png,I0144_01_02.png,7.491250291 431 | I0144_01_05.png,I0144_01_06.png,6.909320269 432 | I0144_01_02.png,I0144_01_06.png,3.471214015 433 | I0144_01_02.png,I0144_01_05.png,6.159935605 434 | I0144_01_02.png,I0144_01_03.png,5.801761174 435 | I0144_01_02.png,I0144_01_04.png,7.770810777 436 | I0145_01_03.png,I0145_01_06.png,5.974976804 437 | I0145_01_04.png,I0145_01_05.png,4.544441434 438 | I0145_01_05.png,I0145_01_06.png,6.002368073 439 | I0145_01_02.png,I0145_01_03.png,4.547261501 440 | I0145_01_01.png,I0145_01_06.png,6.518544884 441 | I0145_01_01.png,I0145_01_04.png,4.18638632 442 | I0145_01_01.png,I0145_01_03.png,5.100032978 443 | I0145_01_04.png,I0145_01_06.png,7.470012868 444 | I0145_01_02.png,I0145_01_05.png,7.912752289 445 | I0145_01_03.png,I0145_01_04.png,5.168263046 446 | I0145_01_01.png,I0145_01_02.png,7.228271866 447 | I0145_01_02.png,I0145_01_04.png,6.821738832 448 | I0145_01_03.png,I0145_01_05.png,5.797705885 449 | I0145_01_01.png,I0145_01_05.png,2.552538412 450 | I0145_01_02.png,I0145_01_06.png,3.901210948 451 | I0149_01_02.png,I0149_01_06.png,5.225460856 452 | I0149_01_02.png,I0149_01_04.png,2.2187398 453 | I0149_01_04.png,I0149_01_06.png,5.960715877 454 | I0149_01_03.png,I0149_01_05.png,2.710666986 455 | I0149_01_01.png,I0149_01_03.png,3.169694819 456 | I0149_01_03.png,I0149_01_06.png,4.067805262 457 | I0149_01_01.png,I0149_01_02.png,3.177752565 458 | I0149_01_05.png,I0149_01_06.png,3.679801527 459 | I0149_01_02.png,I0149_01_05.png,3.054283342 460 | I0149_01_03.png,I0149_01_04.png,3.626336634 461 | I0149_01_01.png,I0149_01_05.png,2.804406769 462 | I0149_01_04.png,I0149_01_05.png,3.251362522 463 | I0149_01_01.png,I0149_01_04.png,2.786691736 464 | I0149_01_01.png,I0149_01_06.png,3.662051012 465 | I0149_01_02.png,I0149_01_03.png,4.292454437 466 | I0150_01_02.png,I0150_01_06.png,4.774008186 467 | I0150_01_02.png,I0150_01_03.png,3.724432892 468 | I0150_01_01.png,I0150_01_03.png,2.484410807 469 | I0150_01_02.png,I0150_01_05.png,4.123247867 470 | I0150_01_01.png,I0150_01_06.png,3.396964254 471 | I0150_01_02.png,I0150_01_04.png,2.997454169 472 | I0150_01_03.png,I0150_01_04.png,4.699603527 473 | I0150_01_03.png,I0150_01_06.png,3.062047419 474 | I0150_01_04.png,I0150_01_05.png,3.837471056 475 | I0150_01_01.png,I0150_01_05.png,2.434880577 476 | I0150_01_04.png,I0150_01_06.png,5.469726833 477 | I0150_01_01.png,I0150_01_04.png,3.365861592 478 | I0150_01_01.png,I0150_01_02.png,2.849987819 479 | I0150_01_05.png,I0150_01_06.png,3.129155683 480 | I0150_01_03.png,I0150_01_05.png,2.489100811 481 | I0151_01_01.png,I0151_01_02.png,2.331246072 482 | I0151_01_04.png,I0151_01_06.png,4.620755628 483 | I0151_01_02.png,I0151_01_04.png,2.995376078 484 | I0151_01_01.png,I0151_01_04.png,2.677560586 485 | I0151_01_02.png,I0151_01_05.png,2.907321334 486 | I0151_01_01.png,I0151_01_03.png,3.659304357 487 | I0151_01_01.png,I0151_01_06.png,3.74368272 488 | I0151_01_02.png,I0151_01_06.png,3.388568934 489 | I0151_01_05.png,I0151_01_06.png,4.191344346 490 | I0151_01_02.png,I0151_01_03.png,2.780012886 491 | I0151_01_03.png,I0151_01_06.png,2.92686699 492 | I0151_01_03.png,I0151_01_04.png,4.444938304 493 | I0151_01_03.png,I0151_01_05.png,3.886414715 494 | I0151_01_04.png,I0151_01_05.png,3.062990593 495 | I0151_01_01.png,I0151_01_05.png,2.377050213 496 | I0152_01_02.png,I0152_01_04.png,2.606766339 497 | I0152_01_03.png,I0152_01_06.png,4.475139134 498 | I0152_01_02.png,I0152_01_06.png,4.36875306 499 | I0152_01_01.png,I0152_01_04.png,3.098346783 500 | I0152_01_01.png,I0152_01_02.png,2.500069182 501 | I0152_01_03.png,I0152_01_05.png,2.712427224 502 | I0152_01_05.png,I0152_01_06.png,4.447889642 503 | I0152_01_04.png,I0152_01_06.png,5.134848945 504 | I0152_01_01.png,I0152_01_05.png,3.083249499 505 | I0152_01_02.png,I0152_01_03.png,3.462379213 506 | I0152_01_02.png,I0152_01_05.png,2.099514106 507 | I0152_01_01.png,I0152_01_03.png,3.953723838 508 | I0152_01_04.png,I0152_01_05.png,2.750487132 509 | I0152_01_03.png,I0152_01_04.png,2.279259842 510 | I0152_01_01.png,I0152_01_06.png,6.785505032 511 | I0156_01_02.png,I0156_01_05.png,7.051883425 512 | I0156_01_04.png,I0156_01_06.png,6.470755111 513 | I0156_01_02.png,I0156_01_03.png,6.501929373 514 | I0156_01_01.png,I0156_01_06.png,6.108850762 515 | I0156_01_01.png,I0156_01_03.png,3.831818793 516 | I0156_01_01.png,I0156_01_05.png,5.224030494 517 | I0156_01_04.png,I0156_01_05.png,4.346183466 518 | I0156_01_03.png,I0156_01_05.png,3.696998514 519 | I0156_01_02.png,I0156_01_04.png,4.102936984 520 | I0156_01_03.png,I0156_01_06.png,5.562994357 521 | I0156_01_02.png,I0156_01_06.png,7.557455599 522 | I0156_01_05.png,I0156_01_06.png,6.541549066 523 | I0156_01_01.png,I0156_01_02.png,4.116364933 524 | I0156_01_01.png,I0156_01_04.png,3.42983373 525 | I0156_01_03.png,I0156_01_04.png,5.019319258 526 | I0162_01_02.png,I0162_01_03.png,2.682509936 527 | I0162_01_03.png,I0162_01_06.png,1.68139866 528 | I0162_01_03.png,I0162_01_05.png,2.626784249 529 | I0162_01_01.png,I0162_01_03.png,3.529751706 530 | I0162_01_01.png,I0162_01_06.png,2.229215393 531 | I0162_01_01.png,I0162_01_05.png,3.343843351 532 | I0162_01_04.png,I0162_01_06.png,2.381807855 533 | I0162_01_04.png,I0162_01_05.png,3.194052332 534 | I0162_01_02.png,I0162_01_05.png,3.938588949 535 | I0162_01_01.png,I0162_01_04.png,3.021964031 536 | I0162_01_05.png,I0162_01_06.png,2.501023538 537 | I0162_01_03.png,I0162_01_04.png,2.110000238 538 | I0162_01_01.png,I0162_01_02.png,4.477966284 539 | I0162_01_02.png,I0162_01_04.png,2.023210628 540 | I0162_01_02.png,I0162_01_06.png,2.785324695 541 | I0165_01_05.png,I0165_01_06.png,3.391258709 542 | I0165_01_02.png,I0165_01_04.png,4.541448823 543 | I0165_01_02.png,I0165_01_03.png,4.706163701 544 | I0165_01_03.png,I0165_01_06.png,3.471452075 545 | I0165_01_02.png,I0165_01_06.png,6.10286878 546 | I0165_01_01.png,I0165_01_03.png,4.01362023 547 | I0165_01_03.png,I0165_01_04.png,6.996582281 548 | I0165_01_02.png,I0165_01_05.png,5.335200321 549 | I0165_01_04.png,I0165_01_05.png,8.240033181 550 | I0165_01_04.png,I0165_01_06.png,9.743219087 551 | I0165_01_01.png,I0165_01_05.png,3.065264229 552 | I0165_01_01.png,I0165_01_04.png,8.492161596 553 | I0165_01_01.png,I0165_01_02.png,5.794719467 554 | I0165_01_03.png,I0165_01_05.png,2.16104628 555 | I0165_01_01.png,I0165_01_06.png,5.392991778 556 | I0167_01_01.png,I0167_01_03.png,3.512461196 557 | I0167_01_02.png,I0167_01_05.png,5.279854653 558 | I0167_01_04.png,I0167_01_06.png,6.267328232 559 | I0167_01_01.png,I0167_01_05.png,5.309423479 560 | I0167_01_02.png,I0167_01_03.png,4.910773444 561 | I0167_01_02.png,I0167_01_04.png,4.169418987 562 | I0167_01_01.png,I0167_01_06.png,4.096585799 563 | I0167_01_05.png,I0167_01_06.png,5.403296116 564 | I0167_01_01.png,I0167_01_04.png,4.835577405 565 | I0167_01_03.png,I0167_01_05.png,5.302226107 566 | I0167_01_04.png,I0167_01_05.png,6.459552875 567 | I0167_01_01.png,I0167_01_02.png,5.586389486 568 | I0167_01_03.png,I0167_01_06.png,3.033346398 569 | I0167_01_02.png,I0167_01_06.png,6.742335453 570 | I0167_01_03.png,I0167_01_04.png,6.215106493 571 | I0169_01_02.png,I0169_01_05.png,5.989457835 572 | I0169_01_04.png,I0169_01_05.png,2.534465853 573 | I0169_01_03.png,I0169_01_04.png,3.418789063 574 | I0169_01_01.png,I0169_01_05.png,5.514513027 575 | I0169_01_01.png,I0169_01_04.png,6.205200623 576 | I0169_01_02.png,I0169_01_04.png,5.04724434 577 | I0169_01_05.png,I0169_01_06.png,3.641943918 578 | I0169_01_01.png,I0169_01_02.png,6.432674093 579 | I0169_01_03.png,I0169_01_05.png,3.045235985 580 | I0169_01_01.png,I0169_01_06.png,3.731535031 581 | I0169_01_02.png,I0169_01_03.png,7.169411156 582 | I0169_01_01.png,I0169_01_03.png,6.738631055 583 | I0169_01_03.png,I0169_01_06.png,6.31068181 584 | I0169_01_02.png,I0169_01_06.png,5.078304771 585 | I0169_01_04.png,I0169_01_06.png,5.063104732 586 | I0172_01_01.png,I0172_01_05.png,2.924352992 587 | I0172_01_01.png,I0172_01_04.png,7.832321221 588 | I0172_01_03.png,I0172_01_06.png,4.870673952 589 | I0172_01_04.png,I0172_01_06.png,6.489285178 590 | I0172_01_01.png,I0172_01_02.png,8.935584197 591 | I0172_01_02.png,I0172_01_03.png,4.227035085 592 | I0172_01_02.png,I0172_01_06.png,7.107614387 593 | I0172_01_01.png,I0172_01_06.png,3.904101497 594 | I0172_01_04.png,I0172_01_05.png,5.744293763 595 | I0172_01_05.png,I0172_01_06.png,3.374495923 596 | I0172_01_02.png,I0172_01_04.png,3.431557359 597 | I0172_01_03.png,I0172_01_04.png,5.105193855 598 | I0172_01_03.png,I0172_01_05.png,3.83513884 599 | I0172_01_01.png,I0172_01_03.png,6.757034373 600 | I0172_01_02.png,I0172_01_05.png,7.335248375 601 | I0179_01_05.png,I0179_01_06.png,2.446332999 602 | I0179_01_02.png,I0179_01_04.png,3.592741783 603 | I0179_01_01.png,I0179_01_05.png,2.554836009 604 | I0179_01_04.png,I0179_01_06.png,4.068424069 605 | I0179_01_03.png,I0179_01_04.png,5.47516312 606 | I0179_01_01.png,I0179_01_04.png,4.759407764 607 | I0179_01_03.png,I0179_01_06.png,2.719817419 608 | I0179_01_01.png,I0179_01_02.png,3.749970529 609 | I0179_01_04.png,I0179_01_05.png,4.751348536 610 | I0179_01_01.png,I0179_01_03.png,3.073132878 611 | I0179_01_01.png,I0179_01_06.png,2.659545048 612 | I0179_01_02.png,I0179_01_03.png,4.461311671 613 | I0179_01_02.png,I0179_01_05.png,3.174665825 614 | I0179_01_02.png,I0179_01_06.png,3.418774652 615 | I0179_01_03.png,I0179_01_05.png,3.261796373 616 | I0183_01_02.png,I0183_01_05.png,4.720986239 617 | I0183_01_04.png,I0183_01_05.png,3.179812338 618 | I0183_01_01.png,I0183_01_05.png,3.519063957 619 | I0183_01_01.png,I0183_01_06.png,4.308301742 620 | I0183_01_04.png,I0183_01_06.png,2.795576206 621 | I0183_01_05.png,I0183_01_06.png,3.610390183 622 | I0183_01_01.png,I0183_01_04.png,2.390191702 623 | I0183_01_01.png,I0183_01_03.png,3.062256607 624 | I0183_01_02.png,I0183_01_03.png,5.49704373 625 | I0183_01_02.png,I0183_01_04.png,3.48573899 626 | I0183_01_02.png,I0183_01_06.png,6.241492929 627 | I0183_01_01.png,I0183_01_02.png,2.95701798 628 | I0183_01_03.png,I0183_01_05.png,3.890571972 629 | I0183_01_03.png,I0183_01_04.png,2.211705557 630 | I0183_01_03.png,I0183_01_06.png,2.951137212 631 | I0185_01_03.png,I0185_01_06.png,3.696026476 632 | I0185_01_01.png,I0185_01_05.png,4.231455084 633 | I0185_01_04.png,I0185_01_06.png,4.813174844 634 | I0185_01_02.png,I0185_01_06.png,4.603897413 635 | I0185_01_01.png,I0185_01_03.png,4.200600562 636 | I0185_01_02.png,I0185_01_05.png,5.146559005 637 | I0185_01_04.png,I0185_01_05.png,4.62658275 638 | I0185_01_03.png,I0185_01_05.png,4.350162914 639 | I0185_01_02.png,I0185_01_04.png,2.249121355 640 | I0185_01_03.png,I0185_01_04.png,3.193680253 641 | I0185_01_01.png,I0185_01_02.png,5.027702216 642 | I0185_01_01.png,I0185_01_04.png,3.841457167 643 | I0185_01_01.png,I0185_01_06.png,4.265632046 644 | I0185_01_05.png,I0185_01_06.png,3.193705646 645 | I0185_01_02.png,I0185_01_03.png,3.314834165 646 | I0194_01_02.png,I0194_01_05.png,11.71634155 647 | I0194_01_03.png,I0194_01_05.png,3.434154404 648 | I0194_01_05.png,I0194_01_06.png,8.825176922 649 | I0194_01_02.png,I0194_01_03.png,13.02282744 650 | I0194_01_01.png,I0194_01_06.png,7.912038353 651 | I0194_01_03.png,I0194_01_04.png,4.075979828 652 | I0194_01_02.png,I0194_01_06.png,4.398848074 653 | I0194_01_01.png,I0194_01_02.png,11.94665995 654 | I0194_01_04.png,I0194_01_06.png,10.31765986 655 | I0194_01_02.png,I0194_01_04.png,15.38252775 656 | I0194_01_01.png,I0194_01_03.png,4.160745214 657 | I0194_01_01.png,I0194_01_05.png,2.899942923 658 | I0194_01_01.png,I0194_01_04.png,3.330463256 659 | I0194_01_04.png,I0194_01_05.png,3.737554071 660 | I0194_01_03.png,I0194_01_06.png,7.814612535 661 | I0198_01_01.png,I0198_01_02.png,5.060020238 662 | I0198_01_05.png,I0198_01_06.png,4.764155417 663 | I0198_01_03.png,I0198_01_05.png,3.775069944 664 | I0198_01_02.png,I0198_01_04.png,5.97391678 665 | I0198_01_02.png,I0198_01_05.png,3.925510518 666 | I0198_01_01.png,I0198_01_06.png,3.060576995 667 | I0198_01_02.png,I0198_01_06.png,5.661828465 668 | I0198_01_04.png,I0198_01_06.png,9.379145179 669 | I0198_01_02.png,I0198_01_03.png,5.44915892 670 | I0198_01_01.png,I0198_01_03.png,4.203465121 671 | I0198_01_03.png,I0198_01_04.png,7.549624286 672 | I0198_01_04.png,I0198_01_05.png,6.75728686 673 | I0198_01_03.png,I0198_01_06.png,4.559890463 674 | I0198_01_01.png,I0198_01_05.png,3.598134838 675 | I0198_01_01.png,I0198_01_04.png,8.034774946 676 | I0199_01_03.png,I0199_01_06.png,5.535668464 677 | I0199_01_02.png,I0199_01_03.png,5.87145682 678 | I0199_01_04.png,I0199_01_05.png,4.749407606 679 | I0199_01_01.png,I0199_01_02.png,3.184525609 680 | I0199_01_02.png,I0199_01_06.png,4.793351706 681 | I0199_01_02.png,I0199_01_05.png,5.096805855 682 | I0199_01_03.png,I0199_01_04.png,6.541077937 683 | I0199_01_01.png,I0199_01_03.png,6.775788704 684 | I0199_01_01.png,I0199_01_05.png,5.075223547 685 | I0199_01_02.png,I0199_01_04.png,3.585073307 686 | I0199_01_05.png,I0199_01_06.png,3.151740246 687 | I0199_01_03.png,I0199_01_05.png,4.349002631 688 | I0199_01_01.png,I0199_01_04.png,2.417706078 689 | I0199_01_01.png,I0199_01_06.png,4.69997468 690 | I0199_01_04.png,I0199_01_06.png,5.173003492 691 | I0203_01_02.png,I0203_01_06.png,3.423761158 692 | I0203_01_01.png,I0203_01_02.png,4.090865683 693 | I0203_01_03.png,I0203_01_06.png,6.45883532 694 | I0203_01_01.png,I0203_01_04.png,2.752043437 695 | I0203_01_02.png,I0203_01_05.png,4.128219296 696 | I0203_01_03.png,I0203_01_04.png,3.748583993 697 | I0203_01_02.png,I0203_01_03.png,4.833020447 698 | I0203_01_03.png,I0203_01_05.png,4.551593394 699 | I0203_01_04.png,I0203_01_06.png,5.580347152 700 | I0203_01_02.png,I0203_01_04.png,4.806560484 701 | I0203_01_04.png,I0203_01_05.png,5.187904367 702 | I0203_01_01.png,I0203_01_05.png,4.495497398 703 | I0203_01_01.png,I0203_01_06.png,3.687286737 704 | I0203_01_01.png,I0203_01_03.png,5.137867336 705 | I0203_01_05.png,I0203_01_06.png,3.033592283 706 | I0209_01_04.png,I0209_01_06.png,6.67392081 707 | I0209_01_03.png,I0209_01_05.png,4.270390971 708 | I0209_01_03.png,I0209_01_04.png,6.663157205 709 | I0209_01_01.png,I0209_01_03.png,3.128141962 710 | I0209_01_01.png,I0209_01_05.png,3.426878424 711 | I0209_01_01.png,I0209_01_02.png,5.614489863 712 | I0209_01_02.png,I0209_01_04.png,7.566865441 713 | I0209_01_04.png,I0209_01_05.png,5.247636513 714 | I0209_01_03.png,I0209_01_06.png,3.461185546 715 | I0209_01_05.png,I0209_01_06.png,4.40533665 716 | I0209_01_01.png,I0209_01_04.png,4.225860416 717 | I0209_01_02.png,I0209_01_06.png,3.897519913 718 | I0209_01_01.png,I0209_01_06.png,4.107465454 719 | I0209_01_02.png,I0209_01_05.png,6.31512541 720 | I0209_01_02.png,I0209_01_03.png,4.278101517 721 | I0210_01_01.png,I0210_01_03.png,3.12189778 722 | I0210_01_01.png,I0210_01_02.png,5.07744926 723 | I0210_01_04.png,I0210_01_06.png,5.667963998 724 | I0210_01_04.png,I0210_01_05.png,3.831669855 725 | I0210_01_02.png,I0210_01_04.png,6.486304649 726 | I0210_01_01.png,I0210_01_06.png,4.213917344 727 | I0210_01_01.png,I0210_01_05.png,4.716071573 728 | I0210_01_03.png,I0210_01_04.png,5.360601563 729 | I0210_01_03.png,I0210_01_06.png,3.454909721 730 | I0210_01_05.png,I0210_01_06.png,5.719879838 731 | I0210_01_02.png,I0210_01_03.png,5.140484174 732 | I0210_01_03.png,I0210_01_05.png,4.422735325 733 | I0210_01_02.png,I0210_01_05.png,6.549875485 734 | I0210_01_02.png,I0210_01_06.png,3.202607905 735 | I0210_01_01.png,I0210_01_04.png,3.715046026 736 | I0214_01_01.png,I0214_01_06.png,3.10896572 737 | I0214_01_03.png,I0214_01_05.png,2.574475124 738 | I0214_01_02.png,I0214_01_04.png,4.947379997 739 | I0214_01_05.png,I0214_01_06.png,3.064752289 740 | I0214_01_02.png,I0214_01_06.png,3.425661375 741 | I0214_01_01.png,I0214_01_04.png,3.180727625 742 | I0214_01_03.png,I0214_01_06.png,3.49562093 743 | I0214_01_01.png,I0214_01_05.png,2.330320383 744 | I0214_01_01.png,I0214_01_03.png,3.093327008 745 | I0214_01_04.png,I0214_01_05.png,3.358857266 746 | I0214_01_03.png,I0214_01_04.png,4.363924515 747 | I0214_01_01.png,I0214_01_02.png,2.900809618 748 | I0214_01_04.png,I0214_01_06.png,6.278126586 749 | I0214_01_02.png,I0214_01_05.png,2.623468334 750 | I0214_01_02.png,I0214_01_03.png,3.456790172 751 | I0215_01_05.png,I0215_01_06.png,6.291059583 752 | I0215_01_01.png,I0215_01_02.png,3.153494382 753 | I0215_01_01.png,I0215_01_05.png,5.831138606 754 | I0215_01_04.png,I0215_01_06.png,6.519769319 755 | I0215_01_02.png,I0215_01_06.png,5.642954705 756 | I0215_01_03.png,I0215_01_04.png,2.978670748 757 | I0215_01_01.png,I0215_01_04.png,4.720486824 758 | I0215_01_02.png,I0215_01_03.png,3.218765481 759 | I0215_01_02.png,I0215_01_04.png,2.582948382 760 | I0215_01_02.png,I0215_01_05.png,2.470757443 761 | I0215_01_01.png,I0215_01_06.png,5.748172206 762 | I0215_01_03.png,I0215_01_05.png,2.760142962 763 | I0215_01_04.png,I0215_01_05.png,2.426408397 764 | I0215_01_03.png,I0215_01_06.png,6.009334856 765 | I0215_01_01.png,I0215_01_03.png,3.784586966 766 | I0216_01_05.png,I0216_01_06.png,5.150738911 767 | I0216_01_02.png,I0216_01_05.png,4.174577301 768 | I0216_01_02.png,I0216_01_04.png,5.925961286 769 | I0216_01_01.png,I0216_01_03.png,2.951553054 770 | I0216_01_03.png,I0216_01_05.png,2.791151538 771 | I0216_01_02.png,I0216_01_06.png,3.757773581 772 | I0216_01_04.png,I0216_01_05.png,6.191688993 773 | I0216_01_01.png,I0216_01_04.png,3.952828136 774 | I0216_01_03.png,I0216_01_04.png,5.097441543 775 | I0216_01_02.png,I0216_01_03.png,2.659840246 776 | I0216_01_04.png,I0216_01_06.png,8.90684284 777 | I0216_01_01.png,I0216_01_06.png,5.196980352 778 | I0216_01_03.png,I0216_01_06.png,6.048994382 779 | I0216_01_01.png,I0216_01_02.png,2.551053149 780 | I0216_01_01.png,I0216_01_05.png,3.761343615 781 | I0217_01_05.png,I0217_01_06.png,5.819008662 782 | I0217_01_02.png,I0217_01_03.png,6.463265302 783 | I0217_01_03.png,I0217_01_05.png,3.740555821 784 | I0217_01_01.png,I0217_01_06.png,6.879160238 785 | I0217_01_02.png,I0217_01_06.png,9.213186021 786 | I0217_01_02.png,I0217_01_05.png,5.992484155 787 | I0217_01_01.png,I0217_01_05.png,3.10192902 788 | I0217_01_03.png,I0217_01_06.png,6.605901531 789 | I0217_01_02.png,I0217_01_04.png,3.516599473 790 | I0217_01_01.png,I0217_01_04.png,4.486518475 791 | I0217_01_04.png,I0217_01_05.png,4.841668114 792 | I0217_01_04.png,I0217_01_06.png,9.371506289 793 | I0217_01_01.png,I0217_01_02.png,5.291451639 794 | I0217_01_03.png,I0217_01_04.png,3.536804464 795 | I0217_01_01.png,I0217_01_03.png,3.915210806 796 | I0222_01_01.png,I0222_01_04.png,3.214573185 797 | I0222_01_03.png,I0222_01_06.png,11.0658228 798 | I0222_01_04.png,I0222_01_05.png,3.890361307 799 | I0222_01_02.png,I0222_01_06.png,8.811313874 800 | I0222_01_02.png,I0222_01_04.png,4.942134879 801 | I0222_01_03.png,I0222_01_04.png,2.613657101 802 | I0222_01_04.png,I0222_01_06.png,10.36368354 803 | I0222_01_01.png,I0222_01_02.png,4.045682269 804 | I0222_01_02.png,I0222_01_05.png,3.252688874 805 | I0222_01_03.png,I0222_01_05.png,2.21382335 806 | I0222_01_01.png,I0222_01_06.png,11.28813453 807 | I0222_01_01.png,I0222_01_05.png,2.482800687 808 | I0222_01_01.png,I0222_01_03.png,2.14234283 809 | I0222_01_05.png,I0222_01_06.png,10.92165182 810 | I0222_01_02.png,I0222_01_03.png,3.734552282 811 | I0225_01_02.png,I0225_01_06.png,7.910678923 812 | I0225_01_01.png,I0225_01_02.png,6.120011986 813 | I0225_01_02.png,I0225_01_05.png,4.930179443 814 | I0225_01_03.png,I0225_01_05.png,1.769615889 815 | I0225_01_03.png,I0225_01_04.png,3.946644157 816 | I0225_01_04.png,I0225_01_06.png,7.744457822 817 | I0225_01_01.png,I0225_01_06.png,5.296195273 818 | I0225_01_05.png,I0225_01_06.png,4.082231808 819 | I0225_01_04.png,I0225_01_05.png,3.369837108 820 | I0225_01_01.png,I0225_01_03.png,3.384836606 821 | I0225_01_03.png,I0225_01_06.png,4.748274879 822 | I0225_01_02.png,I0225_01_03.png,4.429315542 823 | I0225_01_01.png,I0225_01_04.png,3.523905335 824 | I0225_01_01.png,I0225_01_05.png,3.597830838 825 | I0225_01_02.png,I0225_01_04.png,4.5530443 826 | I0236_01_03.png,I0236_01_06.png,7.05976705 827 | I0236_01_01.png,I0236_01_06.png,9.757936841 828 | I0236_01_01.png,I0236_01_03.png,4.647892215 829 | I0236_01_01.png,I0236_01_02.png,4.396897724 830 | I0236_01_02.png,I0236_01_04.png,4.491274907 831 | I0236_01_04.png,I0236_01_06.png,10.0113619 832 | I0236_01_02.png,I0236_01_03.png,3.166676103 833 | I0236_01_05.png,I0236_01_06.png,9.30291561 834 | I0236_01_01.png,I0236_01_04.png,5.035163992 835 | I0236_01_01.png,I0236_01_05.png,4.141014854 836 | I0236_01_03.png,I0236_01_04.png,5.441962932 837 | I0236_01_04.png,I0236_01_05.png,3.562360362 838 | I0236_01_02.png,I0236_01_05.png,2.428171531 839 | I0236_01_03.png,I0236_01_05.png,3.823229881 840 | I0236_01_02.png,I0236_01_06.png,8.099823346 841 | I0243_01_05.png,I0243_01_06.png,5.765880221 842 | I0243_01_02.png,I0243_01_05.png,5.947342345 843 | I0243_01_02.png,I0243_01_04.png,8.094871056 844 | I0243_01_03.png,I0243_01_04.png,7.588702259 845 | I0243_01_02.png,I0243_01_06.png,3.705980957 846 | I0243_01_04.png,I0243_01_05.png,7.355136793 847 | I0243_01_03.png,I0243_01_05.png,2.295881015 848 | I0243_01_01.png,I0243_01_02.png,6.443837949 849 | I0243_01_01.png,I0243_01_06.png,7.79187962 850 | I0243_01_01.png,I0243_01_05.png,3.375796374 851 | I0243_01_01.png,I0243_01_04.png,9.307723766 852 | I0243_01_04.png,I0243_01_06.png,8.030938727 853 | I0243_01_03.png,I0243_01_06.png,7.008324423 854 | I0243_01_02.png,I0243_01_03.png,6.19589884 855 | I0243_01_01.png,I0243_01_03.png,3.27225744 856 | I0246_01_02.png,I0246_01_05.png,5.962736764 857 | I0246_01_01.png,I0246_01_02.png,3.95248723 858 | I0246_01_01.png,I0246_01_03.png,3.314035488 859 | I0246_01_01.png,I0246_01_05.png,4.736920889 860 | I0246_01_05.png,I0246_01_06.png,4.168790088 861 | I0246_01_04.png,I0246_01_05.png,3.051863687 862 | I0246_01_03.png,I0246_01_05.png,2.24108942 863 | I0246_01_02.png,I0246_01_04.png,5.404441521 864 | I0246_01_03.png,I0246_01_06.png,3.412435601 865 | I0246_01_02.png,I0246_01_03.png,5.731041081 866 | I0246_01_04.png,I0246_01_06.png,5.843454612 867 | I0246_01_01.png,I0246_01_06.png,3.922662231 868 | I0246_01_01.png,I0246_01_04.png,3.323205352 869 | I0246_01_02.png,I0246_01_06.png,6.020213096 870 | I0246_01_03.png,I0246_01_04.png,2.039247285 871 | I0254_01_04.png,I0254_01_06.png,8.505367133 872 | I0254_01_03.png,I0254_01_05.png,3.580755365 873 | I0254_01_01.png,I0254_01_03.png,3.618726707 874 | I0254_01_02.png,I0254_01_06.png,5.703374289 875 | I0254_01_01.png,I0254_01_04.png,3.392163044 876 | I0254_01_05.png,I0254_01_06.png,5.314272283 877 | I0254_01_01.png,I0254_01_02.png,2.871296457 878 | I0254_01_01.png,I0254_01_05.png,4.592169294 879 | I0254_01_02.png,I0254_01_03.png,4.252648338 880 | I0254_01_03.png,I0254_01_04.png,4.879248512 881 | I0254_01_04.png,I0254_01_05.png,5.110034509 882 | I0254_01_02.png,I0254_01_05.png,4.638338074 883 | I0254_01_03.png,I0254_01_06.png,6.137039269 884 | I0254_01_01.png,I0254_01_06.png,6.59162791 885 | I0254_01_02.png,I0254_01_04.png,3.744298771 886 | I0255_01_02.png,I0255_01_04.png,3.728055426 887 | I0255_01_02.png,I0255_01_05.png,2.585882311 888 | I0255_01_01.png,I0255_01_05.png,3.912239596 889 | I0255_01_03.png,I0255_01_06.png,6.225818004 890 | I0255_01_01.png,I0255_01_06.png,6.756906222 891 | I0255_01_01.png,I0255_01_03.png,3.172961783 892 | I0255_01_04.png,I0255_01_06.png,8.084222697 893 | I0255_01_02.png,I0255_01_06.png,5.088723592 894 | I0255_01_01.png,I0255_01_04.png,3.52321937 895 | I0255_01_02.png,I0255_01_03.png,2.408148543 896 | I0255_01_01.png,I0255_01_02.png,3.211390578 897 | I0255_01_03.png,I0255_01_04.png,3.489241279 898 | I0255_01_05.png,I0255_01_06.png,4.729433955 899 | I0255_01_04.png,I0255_01_05.png,4.661901799 900 | I0255_01_03.png,I0255_01_05.png,2.105921505 901 | I0258_01_01.png,I0258_01_02.png,2.195166456 902 | I0258_01_02.png,I0258_01_04.png,2.418592052 903 | I0258_01_02.png,I0258_01_06.png,7.385027026 904 | I0258_01_01.png,I0258_01_04.png,2.117585241 905 | I0258_01_01.png,I0258_01_05.png,3.267185563 906 | I0258_01_04.png,I0258_01_06.png,6.123130611 907 | I0258_01_05.png,I0258_01_06.png,4.586151922 908 | I0258_01_03.png,I0258_01_06.png,5.537309242 909 | I0258_01_01.png,I0258_01_03.png,3.767533658 910 | I0258_01_01.png,I0258_01_06.png,6.834544671 911 | I0258_01_02.png,I0258_01_03.png,3.913241839 912 | I0258_01_04.png,I0258_01_05.png,2.662303166 913 | I0258_01_02.png,I0258_01_05.png,3.156444577 914 | I0258_01_03.png,I0258_01_05.png,2.854165053 915 | I0258_01_03.png,I0258_01_04.png,2.830236829 916 | I0259_01_01.png,I0259_01_02.png,5.555000313 917 | I0259_01_02.png,I0259_01_04.png,4.588683081 918 | I0259_01_03.png,I0259_01_04.png,3.609901639 919 | I0259_01_03.png,I0259_01_05.png,2.31183552 920 | I0259_01_04.png,I0259_01_05.png,2.806213905 921 | I0259_01_05.png,I0259_01_06.png,1.75507349 922 | I0259_01_02.png,I0259_01_03.png,7.473201377 923 | I0259_01_02.png,I0259_01_05.png,7.190940957 924 | I0259_01_01.png,I0259_01_05.png,3.085197897 925 | I0259_01_01.png,I0259_01_04.png,2.188345049 926 | I0259_01_04.png,I0259_01_06.png,3.229253391 927 | I0259_01_01.png,I0259_01_03.png,3.048881787 928 | I0259_01_03.png,I0259_01_06.png,1.723167908 929 | I0259_01_02.png,I0259_01_06.png,6.558826103 930 | I0259_01_01.png,I0259_01_06.png,2.085772597 931 | I0261_01_04.png,I0261_01_05.png,2.71109453 932 | I0261_01_03.png,I0261_01_04.png,3.453314115 933 | I0261_01_01.png,I0261_01_04.png,2.648462411 934 | I0261_01_01.png,I0261_01_05.png,1.559329119 935 | I0261_01_03.png,I0261_01_06.png,3.557812329 936 | I0261_01_02.png,I0261_01_05.png,3.133704292 937 | I0261_01_01.png,I0261_01_06.png,3.412713801 938 | I0261_01_02.png,I0261_01_06.png,4.358633009 939 | I0261_01_01.png,I0261_01_02.png,2.783847222 940 | I0261_01_02.png,I0261_01_03.png,3.547073684 941 | I0261_01_03.png,I0261_01_05.png,1.997141032 942 | I0261_01_02.png,I0261_01_04.png,1.943943831 943 | I0261_01_04.png,I0261_01_06.png,4.901709063 944 | I0261_01_01.png,I0261_01_03.png,2.236208198 945 | I0261_01_05.png,I0261_01_06.png,3.833357291 946 | I0270_01_03.png,I0270_01_04.png,4.95127215 947 | I0270_01_01.png,I0270_01_05.png,1.99764614 948 | I0270_01_02.png,I0270_01_04.png,4.052324923 949 | I0270_01_02.png,I0270_01_06.png,3.862865045 950 | I0270_01_01.png,I0270_01_04.png,3.493900812 951 | I0270_01_04.png,I0270_01_06.png,4.685469139 952 | I0270_01_04.png,I0270_01_05.png,4.265682042 953 | I0270_01_01.png,I0270_01_02.png,2.580495753 954 | I0270_01_01.png,I0270_01_06.png,3.60084956 955 | I0270_01_02.png,I0270_01_03.png,3.656725135 956 | I0270_01_03.png,I0270_01_05.png,2.374762591 957 | I0270_01_02.png,I0270_01_05.png,3.249171004 958 | I0270_01_01.png,I0270_01_03.png,3.033838283 959 | I0270_01_03.png,I0270_01_06.png,2.694149153 960 | I0270_01_05.png,I0270_01_06.png,2.742733944 961 | I0282_01_02.png,I0282_01_05.png,7.17416265 962 | I0282_01_03.png,I0282_01_06.png,5.727234693 963 | I0282_01_04.png,I0282_01_05.png,6.132373053 964 | I0282_01_01.png,I0282_01_02.png,9.564472038 965 | I0282_01_01.png,I0282_01_05.png,3.480911338 966 | I0282_01_05.png,I0282_01_06.png,4.497768158 967 | I0282_01_01.png,I0282_01_04.png,3.645928797 968 | I0282_01_01.png,I0282_01_03.png,2.98743045 969 | I0282_01_03.png,I0282_01_05.png,3.44510985 970 | I0282_01_03.png,I0282_01_04.png,4.674066091 971 | I0282_01_04.png,I0282_01_06.png,7.223567659 972 | I0282_01_01.png,I0282_01_06.png,6.657513701 973 | I0282_01_02.png,I0282_01_06.png,4.587770669 974 | I0282_01_02.png,I0282_01_03.png,9.978005321 975 | I0282_01_02.png,I0282_01_04.png,9.884921783 976 | I0283_01_03.png,I0283_01_04.png,7.302767658 977 | I0283_01_04.png,I0283_01_05.png,10.00955488 978 | I0283_01_01.png,I0283_01_03.png,3.095951212 979 | I0283_01_02.png,I0283_01_03.png,3.493113438 980 | I0283_01_02.png,I0283_01_05.png,6.12271075 981 | I0283_01_03.png,I0283_01_05.png,5.673086142 982 | I0283_01_01.png,I0283_01_02.png,2.906796414 983 | I0283_01_03.png,I0283_01_06.png,6.232833521 984 | I0283_01_02.png,I0283_01_04.png,7.034807397 985 | I0283_01_05.png,I0283_01_06.png,3.575607805 986 | I0283_01_02.png,I0283_01_06.png,7.200035691 987 | I0283_01_01.png,I0283_01_06.png,5.917051807 988 | I0283_01_04.png,I0283_01_06.png,9.815228189 989 | I0283_01_01.png,I0283_01_05.png,5.615742577 990 | I0283_01_01.png,I0283_01_04.png,8.378830862 991 | I0297_01_05.png,I0297_01_06.png,7.894004306 992 | I0297_01_02.png,I0297_01_03.png,3.319165716 993 | I0297_01_03.png,I0297_01_04.png,2.942662599 994 | I0297_01_04.png,I0297_01_05.png,3.354198615 995 | I0297_01_01.png,I0297_01_02.png,2.6492866 996 | I0297_01_01.png,I0297_01_04.png,3.639229681 997 | I0297_01_03.png,I0297_01_05.png,2.311024599 998 | I0297_01_01.png,I0297_01_03.png,2.620961739 999 | I0297_01_01.png,I0297_01_06.png,8.277361937 1000 | I0297_01_03.png,I0297_01_06.png,8.102341086 1001 | I0297_01_02.png,I0297_01_04.png,4.703702045 1002 | I0297_01_01.png,I0297_01_05.png,2.699530608 1003 | I0297_01_02.png,I0297_01_06.png,5.700349186 1004 | I0297_01_04.png,I0297_01_06.png,9.895896271 1005 | I0297_01_02.png,I0297_01_05.png,3.477354172 1006 | I0308_01_01.png,I0308_01_03.png,3.877895335 1007 | I0308_01_04.png,I0308_01_06.png,5.361031841 1008 | I0308_01_03.png,I0308_01_06.png,4.820289194 1009 | I0308_01_05.png,I0308_01_06.png,3.872520381 1010 | I0308_01_03.png,I0308_01_04.png,3.461355158 1011 | I0308_01_01.png,I0308_01_02.png,5.172166102 1012 | I0308_01_02.png,I0308_01_05.png,3.252996957 1013 | I0308_01_02.png,I0308_01_04.png,3.132070185 1014 | I0308_01_03.png,I0308_01_05.png,2.57115258 1015 | I0308_01_02.png,I0308_01_03.png,3.864736947 1016 | I0308_01_02.png,I0308_01_06.png,5.234270434 1017 | I0308_01_04.png,I0308_01_05.png,3.049265939 1018 | I0308_01_01.png,I0308_01_05.png,5.181785299 1019 | I0308_01_01.png,I0308_01_06.png,5.992549023 1020 | I0308_01_01.png,I0308_01_04.png,4.647985364 1021 | I0317_01_04.png,I0317_01_06.png,5.67320301 1022 | I0317_01_02.png,I0317_01_05.png,3.587394139 1023 | I0317_01_04.png,I0317_01_05.png,8.153852926 1024 | I0317_01_01.png,I0317_01_05.png,2.764166437 1025 | I0317_01_03.png,I0317_01_04.png,10.24186346 1026 | I0317_01_02.png,I0317_01_06.png,4.800286936 1027 | I0317_01_01.png,I0317_01_02.png,3.767140958 1028 | I0317_01_03.png,I0317_01_06.png,5.829499324 1029 | I0317_01_01.png,I0317_01_06.png,6.268644113 1030 | I0317_01_02.png,I0317_01_04.png,8.750911902 1031 | I0317_01_02.png,I0317_01_03.png,4.518100186 1032 | I0317_01_03.png,I0317_01_05.png,2.69593557 1033 | I0317_01_05.png,I0317_01_06.png,3.814237861 1034 | I0317_01_01.png,I0317_01_04.png,9.748340138 1035 | I0317_01_01.png,I0317_01_03.png,2.550245533 1036 | I0318_01_01.png,I0318_01_03.png,3.218497578 1037 | I0318_01_03.png,I0318_01_04.png,5.23782834 1038 | I0318_01_04.png,I0318_01_06.png,4.236010417 1039 | I0318_01_02.png,I0318_01_06.png,2.777828425 1040 | I0318_01_05.png,I0318_01_06.png,2.443305751 1041 | I0318_01_04.png,I0318_01_05.png,3.950765315 1042 | I0318_01_02.png,I0318_01_03.png,3.153302391 1043 | I0318_01_02.png,I0318_01_05.png,2.65917738 1044 | I0318_01_01.png,I0318_01_02.png,2.716163691 1045 | I0318_01_03.png,I0318_01_06.png,2.594578018 1046 | I0318_01_01.png,I0318_01_05.png,1.895832025 1047 | I0318_01_03.png,I0318_01_05.png,2.332644461 1048 | I0318_01_01.png,I0318_01_04.png,4.356412867 1049 | I0318_01_02.png,I0318_01_04.png,3.644986939 1050 | I0318_01_01.png,I0318_01_06.png,2.20477181 1051 | I0335_01_01.png,I0335_01_02.png,4.972263189 1052 | I0335_01_02.png,I0335_01_06.png,4.858741541 1053 | I0335_01_01.png,I0335_01_06.png,5.764150898 1054 | I0335_01_02.png,I0335_01_03.png,4.086367081 1055 | I0335_01_03.png,I0335_01_04.png,2.860661564 1056 | I0335_01_03.png,I0335_01_05.png,3.798181854 1057 | I0335_01_04.png,I0335_01_06.png,5.897230899 1058 | I0335_01_01.png,I0335_01_03.png,3.146756508 1059 | I0335_01_05.png,I0335_01_06.png,4.763027724 1060 | I0335_01_02.png,I0335_01_04.png,3.984693322 1061 | I0335_01_04.png,I0335_01_05.png,3.749572039 1062 | I0335_01_01.png,I0335_01_05.png,4.317567267 1063 | I0335_01_01.png,I0335_01_04.png,2.77236957 1064 | I0335_01_03.png,I0335_01_06.png,5.008555097 1065 | I0335_01_02.png,I0335_01_05.png,2.420780825 1066 | I0342_01_03.png,I0342_01_04.png,2.3730904 1067 | I0342_01_01.png,I0342_01_05.png,4.494171012 1068 | I0342_01_04.png,I0342_01_06.png,6.010496836 1069 | I0342_01_01.png,I0342_01_03.png,3.227460924 1070 | I0342_01_02.png,I0342_01_04.png,4.482440458 1071 | I0342_01_03.png,I0342_01_06.png,5.641993615 1072 | I0342_01_05.png,I0342_01_06.png,5.307912424 1073 | I0342_01_04.png,I0342_01_05.png,4.335943171 1074 | I0342_01_02.png,I0342_01_05.png,4.937517629 1075 | I0342_01_02.png,I0342_01_03.png,4.645014551 1076 | I0342_01_01.png,I0342_01_02.png,3.756360768 1077 | I0342_01_01.png,I0342_01_04.png,2.287217278 1078 | I0342_01_02.png,I0342_01_06.png,7.500697137 1079 | I0342_01_03.png,I0342_01_05.png,3.74894959 1080 | I0342_01_01.png,I0342_01_06.png,6.005666672 1081 | I0344_01_01.png,I0344_01_02.png,2.286184929 1082 | I0344_01_02.png,I0344_01_06.png,4.797629428 1083 | I0344_01_02.png,I0344_01_05.png,3.186369228 1084 | I0344_01_03.png,I0344_01_05.png,4.443471479 1085 | I0344_01_01.png,I0344_01_03.png,3.694305137 1086 | I0344_01_01.png,I0344_01_04.png,2.44361618 1087 | I0344_01_01.png,I0344_01_05.png,3.26795792 1088 | I0344_01_05.png,I0344_01_06.png,4.661236271 1089 | I0344_01_04.png,I0344_01_06.png,7.081314106 1090 | I0344_01_01.png,I0344_01_06.png,6.768717081 1091 | I0344_01_02.png,I0344_01_04.png,3.515077756 1092 | I0344_01_02.png,I0344_01_03.png,3.70919258 1093 | I0344_01_03.png,I0344_01_04.png,3.059539698 1094 | I0344_01_03.png,I0344_01_06.png,7.617711848 1095 | I0344_01_04.png,I0344_01_05.png,4.540272704 1096 | I0347_01_01.png,I0347_01_03.png,3.367169002 1097 | I0347_01_04.png,I0347_01_06.png,4.447114871 1098 | I0347_01_02.png,I0347_01_04.png,3.141361361 1099 | I0347_01_04.png,I0347_01_05.png,2.456995986 1100 | I0347_01_02.png,I0347_01_06.png,2.881369558 1101 | I0347_01_02.png,I0347_01_05.png,3.367919165 1102 | I0347_01_03.png,I0347_01_04.png,4.745598808 1103 | I0347_01_01.png,I0347_01_06.png,3.015264477 1104 | I0347_01_03.png,I0347_01_06.png,2.646896558 1105 | I0347_01_05.png,I0347_01_06.png,2.983524167 1106 | I0347_01_01.png,I0347_01_02.png,2.983318947 1107 | I0347_01_01.png,I0347_01_04.png,3.021715315 1108 | I0347_01_01.png,I0347_01_05.png,2.197470819 1109 | I0347_01_03.png,I0347_01_05.png,3.637545437 1110 | I0347_01_02.png,I0347_01_03.png,4.609180578 1111 | I0352_01_04.png,I0352_01_06.png,6.979843172 1112 | I0352_01_04.png,I0352_01_05.png,3.540250869 1113 | I0352_01_01.png,I0352_01_05.png,2.361652897 1114 | I0352_01_03.png,I0352_01_04.png,4.024539891 1115 | I0352_01_05.png,I0352_01_06.png,5.31925023 1116 | I0352_01_03.png,I0352_01_06.png,3.997847515 1117 | I0352_01_01.png,I0352_01_03.png,3.949707316 1118 | I0352_01_01.png,I0352_01_04.png,2.143939572 1119 | I0352_01_01.png,I0352_01_02.png,5.586198154 1120 | I0352_01_02.png,I0352_01_06.png,2.528216509 1121 | I0352_01_02.png,I0352_01_04.png,5.339162522 1122 | I0352_01_01.png,I0352_01_06.png,6.701205047 1123 | I0352_01_03.png,I0352_01_05.png,3.063875221 1124 | I0352_01_02.png,I0352_01_03.png,3.823124293 1125 | I0352_01_02.png,I0352_01_05.png,5.255793313 1126 | I0356_01_02.png,I0356_01_03.png,2.918692199 1127 | I0356_01_02.png,I0356_01_04.png,3.152019481 1128 | I0356_01_01.png,I0356_01_05.png,3.189011067 1129 | I0356_01_01.png,I0356_01_04.png,3.744119033 1130 | I0356_01_03.png,I0356_01_05.png,4.413352855 1131 | I0356_01_05.png,I0356_01_06.png,4.785640507 1132 | I0356_01_01.png,I0356_01_03.png,4.595525426 1133 | I0356_01_03.png,I0356_01_06.png,5.049782184 1134 | I0356_01_03.png,I0356_01_04.png,4.042232066 1135 | I0356_01_04.png,I0356_01_05.png,4.706762507 1136 | I0356_01_02.png,I0356_01_06.png,6.537840327 1137 | I0356_01_01.png,I0356_01_02.png,3.83515352 1138 | I0356_01_02.png,I0356_01_05.png,3.756148629 1139 | I0356_01_01.png,I0356_01_06.png,5.021402589 1140 | I0356_01_04.png,I0356_01_06.png,5.968248581 1141 | I0359_01_01.png,I0359_01_06.png,4.241692209 1142 | I0359_01_04.png,I0359_01_05.png,3.799054202 1143 | I0359_01_04.png,I0359_01_06.png,3.927324629 1144 | I0359_01_02.png,I0359_01_06.png,4.673904708 1145 | I0359_01_02.png,I0359_01_05.png,3.051474865 1146 | I0359_01_05.png,I0359_01_06.png,3.487049665 1147 | I0359_01_03.png,I0359_01_04.png,3.739902955 1148 | I0359_01_03.png,I0359_01_06.png,5.078995512 1149 | I0359_01_01.png,I0359_01_04.png,3.507207038 1150 | I0359_01_01.png,I0359_01_02.png,2.822599741 1151 | I0359_01_01.png,I0359_01_03.png,2.792203638 1152 | I0359_01_02.png,I0359_01_03.png,4.315017304 1153 | I0359_01_02.png,I0359_01_04.png,2.82983425 1154 | I0359_01_01.png,I0359_01_05.png,2.193251657 1155 | I0359_01_03.png,I0359_01_05.png,3.074694361 1156 | I0366_01_02.png,I0366_01_03.png,10.15855423 1157 | I0366_01_01.png,I0366_01_04.png,3.28772288 1158 | I0366_01_04.png,I0366_01_05.png,5.862359895 1159 | I0366_01_03.png,I0366_01_05.png,3.930783701 1160 | I0366_01_03.png,I0366_01_06.png,6.433920752 1161 | I0366_01_01.png,I0366_01_02.png,10.44059214 1162 | I0366_01_03.png,I0366_01_04.png,4.826521012 1163 | I0366_01_02.png,I0366_01_06.png,3.48049546 1164 | I0366_01_04.png,I0366_01_06.png,7.232917873 1165 | I0366_01_05.png,I0366_01_06.png,4.256204967 1166 | I0366_01_01.png,I0366_01_03.png,1.982375131 1167 | I0366_01_02.png,I0366_01_05.png,7.287054842 1168 | I0366_01_01.png,I0366_01_05.png,3.754111935 1169 | I0366_01_02.png,I0366_01_04.png,10.5472553 1170 | I0366_01_01.png,I0366_01_06.png,7.649872297 1171 | I0372_01_02.png,I0372_01_03.png,10.81104219 1172 | I0372_01_05.png,I0372_01_06.png,6.144384815 1173 | I0372_01_02.png,I0372_01_06.png,12.22138902 1174 | I0372_01_01.png,I0372_01_06.png,3.664908457 1175 | I0372_01_01.png,I0372_01_03.png,5.711348052 1176 | I0372_01_01.png,I0372_01_05.png,6.525575673 1177 | I0372_01_02.png,I0372_01_04.png,11.02303203 1178 | I0372_01_01.png,I0372_01_04.png,6.995075018 1179 | I0372_01_03.png,I0372_01_05.png,8.189977843 1180 | I0372_01_02.png,I0372_01_05.png,12.22730671 1181 | I0372_01_03.png,I0372_01_04.png,7.232044351 1182 | I0372_01_04.png,I0372_01_06.png,5.725021672 1183 | I0372_01_04.png,I0372_01_05.png,2.439373931 1184 | I0372_01_03.png,I0372_01_06.png,4.137019653 1185 | I0372_01_01.png,I0372_01_02.png,10.01831937 1186 | I0379_01_01.png,I0379_01_02.png,2.771625012 1187 | I0379_01_05.png,I0379_01_06.png,3.547880212 1188 | I0379_01_01.png,I0379_01_06.png,2.517755443 1189 | I0379_01_02.png,I0379_01_05.png,3.631913904 1190 | I0379_01_01.png,I0379_01_03.png,4.453319326 1191 | I0379_01_02.png,I0379_01_04.png,3.610654004 1192 | I0379_01_04.png,I0379_01_06.png,2.702916576 1193 | I0379_01_03.png,I0379_01_04.png,2.885866199 1194 | I0379_01_04.png,I0379_01_05.png,3.273921717 1195 | I0379_01_03.png,I0379_01_06.png,2.931449355 1196 | I0379_01_02.png,I0379_01_06.png,3.25297769 1197 | I0379_01_01.png,I0379_01_05.png,2.547262287 1198 | I0379_01_01.png,I0379_01_04.png,2.897113454 1199 | I0379_01_03.png,I0379_01_05.png,4.667011522 1200 | I0379_01_02.png,I0379_01_03.png,4.820021896 1201 | I0385_01_01.png,I0385_01_03.png,3.841894261 1202 | I0385_01_02.png,I0385_01_03.png,3.112583978 1203 | I0385_01_05.png,I0385_01_06.png,3.27015117 1204 | I0385_01_01.png,I0385_01_05.png,4.790217592 1205 | I0385_01_01.png,I0385_01_02.png,4.92370008 1206 | I0385_01_01.png,I0385_01_06.png,2.657241007 1207 | I0385_01_02.png,I0385_01_06.png,3.236065623 1208 | I0385_01_03.png,I0385_01_06.png,2.987355875 1209 | I0385_01_02.png,I0385_01_04.png,4.319831901 1210 | I0385_01_03.png,I0385_01_04.png,4.120914934 1211 | I0385_01_01.png,I0385_01_04.png,6.345834682 1212 | I0385_01_02.png,I0385_01_05.png,2.506258234 1213 | I0385_01_04.png,I0385_01_05.png,6.350310618 1214 | I0385_01_03.png,I0385_01_05.png,2.64634733 1215 | I0385_01_04.png,I0385_01_06.png,3.641883195 1216 | I0397_01_04.png,I0397_01_06.png,10.68853134 1217 | I0397_01_01.png,I0397_01_03.png,3.298611531 1218 | I0397_01_01.png,I0397_01_05.png,2.611227399 1219 | I0397_01_05.png,I0397_01_06.png,4.900006261 1220 | I0397_01_03.png,I0397_01_06.png,7.465073405 1221 | I0397_01_04.png,I0397_01_05.png,8.126266776 1222 | I0397_01_03.png,I0397_01_04.png,5.318458266 1223 | I0397_01_01.png,I0397_01_04.png,6.327098984 1224 | I0397_01_01.png,I0397_01_06.png,5.827784834 1225 | I0397_01_02.png,I0397_01_06.png,7.781725995 1226 | I0397_01_02.png,I0397_01_05.png,5.488020273 1227 | I0397_01_01.png,I0397_01_02.png,2.443794911 1228 | I0397_01_03.png,I0397_01_05.png,4.162773537 1229 | I0397_01_02.png,I0397_01_03.png,3.409704709 1230 | I0397_01_02.png,I0397_01_04.png,4.407925177 1231 | I0400_01_04.png,I0400_01_06.png,6.196455569 1232 | I0400_01_03.png,I0400_01_04.png,6.375775055 1233 | I0400_01_02.png,I0400_01_03.png,3.798232501 1234 | I0400_01_02.png,I0400_01_06.png,4.12900667 1235 | I0400_01_03.png,I0400_01_06.png,2.170101115 1236 | I0400_01_02.png,I0400_01_05.png,2.962148193 1237 | I0400_01_01.png,I0400_01_03.png,3.845520781 1238 | I0400_01_01.png,I0400_01_02.png,3.104847813 1239 | I0400_01_02.png,I0400_01_04.png,2.856792562 1240 | I0400_01_01.png,I0400_01_05.png,3.364133656 1241 | I0400_01_04.png,I0400_01_05.png,5.780199901 1242 | I0400_01_01.png,I0400_01_06.png,3.721111194 1243 | I0400_01_03.png,I0400_01_05.png,3.154341694 1244 | I0400_01_05.png,I0400_01_06.png,2.487040635 1245 | I0400_01_01.png,I0400_01_04.png,3.017518215 1246 | I0403_01_02.png,I0403_01_04.png,2.764050389 1247 | I0403_01_02.png,I0403_01_06.png,3.324807731 1248 | I0403_01_03.png,I0403_01_04.png,2.389030961 1249 | I0403_01_05.png,I0403_01_06.png,2.560440095 1250 | I0403_01_01.png,I0403_01_05.png,2.070611802 1251 | I0403_01_02.png,I0403_01_03.png,2.914990155 1252 | I0403_01_01.png,I0403_01_04.png,2.801915217 1253 | I0403_01_01.png,I0403_01_02.png,2.496260262 1254 | I0403_01_04.png,I0403_01_06.png,2.683528601 1255 | I0403_01_04.png,I0403_01_05.png,2.266742921 1256 | I0403_01_02.png,I0403_01_05.png,2.62315514 1257 | I0403_01_01.png,I0403_01_03.png,2.12418919 1258 | I0403_01_01.png,I0403_01_06.png,4.146459032 1259 | I0403_01_03.png,I0403_01_05.png,2.303872814 1260 | I0403_01_03.png,I0403_01_06.png,3.00949382 1261 | I0406_01_01.png,I0406_01_02.png,7.237925354 1262 | I0406_01_02.png,I0406_01_04.png,7.731561668 1263 | I0406_01_01.png,I0406_01_06.png,6.493509356 1264 | I0406_01_01.png,I0406_01_04.png,3.333518627 1265 | I0406_01_03.png,I0406_01_04.png,3.770699194 1266 | I0406_01_04.png,I0406_01_06.png,7.703784185 1267 | I0406_01_02.png,I0406_01_06.png,4.845917875 1268 | I0406_01_04.png,I0406_01_05.png,5.86164015 1269 | I0406_01_05.png,I0406_01_06.png,5.763395242 1270 | I0406_01_02.png,I0406_01_05.png,7.340742024 1271 | I0406_01_03.png,I0406_01_05.png,5.605035156 1272 | I0406_01_01.png,I0406_01_05.png,4.845141646 1273 | I0406_01_03.png,I0406_01_06.png,7.807615675 1274 | I0406_01_02.png,I0406_01_03.png,7.910741447 1275 | I0406_01_01.png,I0406_01_03.png,3.64006811 1276 | I0411_01_01.png,I0411_01_04.png,6.21093874 1277 | I0411_01_01.png,I0411_01_02.png,4.199501944 1278 | I0411_01_05.png,I0411_01_06.png,3.976934811 1279 | I0411_01_04.png,I0411_01_05.png,7.192235991 1280 | I0411_01_02.png,I0411_01_05.png,4.450221698 1281 | I0411_01_01.png,I0411_01_06.png,6.95513878 1282 | I0411_01_03.png,I0411_01_06.png,6.781127958 1283 | I0411_01_04.png,I0411_01_06.png,10.05087698 1284 | I0411_01_03.png,I0411_01_05.png,3.678732149 1285 | I0411_01_01.png,I0411_01_05.png,4.225675985 1286 | I0411_01_02.png,I0411_01_04.png,7.309544684 1287 | I0411_01_01.png,I0411_01_03.png,3.17392259 1288 | I0411_01_02.png,I0411_01_06.png,5.975551759 1289 | I0411_01_02.png,I0411_01_03.png,4.638793938 1290 | I0411_01_03.png,I0411_01_04.png,6.055382698 1291 | I0412_01_01.png,I0412_01_04.png,4.542278993 1292 | I0412_01_02.png,I0412_01_04.png,4.986102008 1293 | I0412_01_01.png,I0412_01_05.png,3.904259965 1294 | I0412_01_03.png,I0412_01_06.png,3.506996548 1295 | I0412_01_01.png,I0412_01_03.png,4.641044141 1296 | I0412_01_01.png,I0412_01_06.png,2.546430789 1297 | I0412_01_03.png,I0412_01_05.png,2.948729338 1298 | I0412_01_04.png,I0412_01_06.png,4.558680635 1299 | I0412_01_03.png,I0412_01_04.png,4.779276397 1300 | I0412_01_02.png,I0412_01_06.png,2.91331472 1301 | I0412_01_02.png,I0412_01_03.png,3.148983107 1302 | I0412_01_01.png,I0412_01_02.png,2.831417071 1303 | I0412_01_04.png,I0412_01_05.png,3.573146585 1304 | I0412_01_05.png,I0412_01_06.png,3.443468536 1305 | I0412_01_02.png,I0412_01_05.png,2.777694737 1306 | I0413_01_04.png,I0413_01_06.png,8.980299783 1307 | I0413_01_01.png,I0413_01_06.png,6.380613 1308 | I0413_01_02.png,I0413_01_05.png,2.908313433 1309 | I0413_01_02.png,I0413_01_03.png,4.032803166 1310 | I0413_01_01.png,I0413_01_02.png,4.549941867 1311 | I0413_01_03.png,I0413_01_05.png,2.286717077 1312 | I0413_01_05.png,I0413_01_06.png,5.653744774 1313 | I0413_01_03.png,I0413_01_06.png,5.984052588 1314 | I0413_01_04.png,I0413_01_05.png,5.860864884 1315 | I0413_01_02.png,I0413_01_04.png,5.728169264 1316 | I0413_01_01.png,I0413_01_03.png,4.85512439 1317 | I0413_01_02.png,I0413_01_06.png,5.687918101 1318 | I0413_01_03.png,I0413_01_04.png,6.148926346 1319 | I0413_01_01.png,I0413_01_04.png,6.877651846 1320 | I0413_01_01.png,I0413_01_05.png,4.186540829 1321 | I0420_01_02.png,I0420_01_03.png,4.108658458 1322 | I0420_01_05.png,I0420_01_06.png,4.045286264 1323 | I0420_01_01.png,I0420_01_02.png,3.074296801 1324 | I0420_01_02.png,I0420_01_05.png,4.047568009 1325 | I0420_01_01.png,I0420_01_04.png,4.811897428 1326 | I0420_01_04.png,I0420_01_05.png,2.971448612 1327 | I0420_01_01.png,I0420_01_03.png,3.706855676 1328 | I0420_01_02.png,I0420_01_06.png,3.757696415 1329 | I0420_01_03.png,I0420_01_04.png,2.838360695 1330 | I0420_01_02.png,I0420_01_04.png,3.737199925 1331 | I0420_01_03.png,I0420_01_05.png,3.069917847 1332 | I0420_01_03.png,I0420_01_06.png,3.907550269 1333 | I0420_01_04.png,I0420_01_06.png,3.391615981 1334 | I0420_01_01.png,I0420_01_06.png,3.200185714 1335 | I0420_01_01.png,I0420_01_05.png,4.635541704 1336 | I0423_01_02.png,I0423_01_05.png,2.368823809 1337 | I0423_01_02.png,I0423_01_06.png,1.996885388 1338 | I0423_01_01.png,I0423_01_03.png,4.142409121 1339 | I0423_01_02.png,I0423_01_04.png,1.932770548 1340 | I0423_01_03.png,I0423_01_06.png,4.062302702 1341 | I0423_01_01.png,I0423_01_02.png,3.235573889 1342 | I0423_01_01.png,I0423_01_04.png,4.024468959 1343 | I0423_01_02.png,I0423_01_03.png,3.702057635 1344 | I0423_01_01.png,I0423_01_05.png,3.676968734 1345 | I0423_01_04.png,I0423_01_05.png,2.251898232 1346 | I0423_01_03.png,I0423_01_04.png,4.263242809 1347 | I0423_01_04.png,I0423_01_06.png,1.985641401 1348 | I0423_01_03.png,I0423_01_05.png,2.613472242 1349 | I0423_01_01.png,I0423_01_06.png,3.857665324 1350 | I0423_01_05.png,I0423_01_06.png,3.037823108 1351 | I0428_01_02.png,I0428_01_05.png,4.701602703 1352 | I0428_01_01.png,I0428_01_06.png,3.742119191 1353 | I0428_01_01.png,I0428_01_05.png,4.382298638 1354 | I0428_01_03.png,I0428_01_04.png,4.939893826 1355 | I0428_01_02.png,I0428_01_04.png,6.405143408 1356 | I0428_01_01.png,I0428_01_03.png,5.009061807 1357 | I0428_01_03.png,I0428_01_05.png,4.545461002 1358 | I0428_01_01.png,I0428_01_02.png,5.563894802 1359 | I0428_01_01.png,I0428_01_04.png,4.825634751 1360 | I0428_01_02.png,I0428_01_03.png,2.508341044 1361 | I0428_01_04.png,I0428_01_05.png,2.667598484 1362 | I0428_01_05.png,I0428_01_06.png,2.89029948 1363 | I0428_01_02.png,I0428_01_06.png,4.451769477 1364 | I0428_01_04.png,I0428_01_06.png,2.814065538 1365 | I0428_01_03.png,I0428_01_06.png,3.365657516 1366 | I0437_01_04.png,I0437_01_05.png,4.894215471 1367 | I0437_01_02.png,I0437_01_06.png,3.210520049 1368 | I0437_01_01.png,I0437_01_02.png,3.969925725 1369 | I0437_01_01.png,I0437_01_05.png,4.592343621 1370 | I0437_01_05.png,I0437_01_06.png,2.73374207 1371 | I0437_01_03.png,I0437_01_06.png,3.439594675 1372 | I0437_01_01.png,I0437_01_06.png,2.995656592 1373 | I0437_01_02.png,I0437_01_03.png,2.898257314 1374 | I0437_01_02.png,I0437_01_04.png,4.686684911 1375 | I0437_01_04.png,I0437_01_06.png,3.530172225 1376 | I0437_01_02.png,I0437_01_05.png,3.634961305 1377 | I0437_01_01.png,I0437_01_04.png,2.269362721 1378 | I0437_01_01.png,I0437_01_03.png,4.607924151 1379 | I0437_01_03.png,I0437_01_04.png,5.589212431 1380 | I0437_01_03.png,I0437_01_05.png,4.330082 1381 | I0444_01_04.png,I0444_01_06.png,2.289961034 1382 | I0444_01_02.png,I0444_01_04.png,4.442127038 1383 | I0444_01_02.png,I0444_01_05.png,2.734912373 1384 | I0444_01_03.png,I0444_01_06.png,3.30167428 1385 | I0444_01_02.png,I0444_01_06.png,3.898570234 1386 | I0444_01_01.png,I0444_01_04.png,4.231936602 1387 | I0444_01_01.png,I0444_01_05.png,2.521138149 1388 | I0444_01_02.png,I0444_01_03.png,3.651911213 1389 | I0444_01_05.png,I0444_01_06.png,4.923172651 1390 | I0444_01_01.png,I0444_01_02.png,2.709975458 1391 | I0444_01_01.png,I0444_01_03.png,3.702807034 1392 | I0444_01_04.png,I0444_01_05.png,4.395527616 1393 | I0444_01_03.png,I0444_01_05.png,4.182905388 1394 | I0444_01_03.png,I0444_01_04.png,4.387873457 1395 | I0444_01_01.png,I0444_01_06.png,5.101695722 1396 | I0445_01_01.png,I0445_01_06.png,6.627445782 1397 | I0445_01_01.png,I0445_01_04.png,7.71927116 1398 | I0445_01_05.png,I0445_01_06.png,5.597362716 1399 | I0445_01_03.png,I0445_01_05.png,8.842417079 1400 | I0445_01_03.png,I0445_01_04.png,8.615631472 1401 | I0445_01_04.png,I0445_01_05.png,5.665420745 1402 | I0445_01_02.png,I0445_01_04.png,4.925730124 1403 | I0445_01_04.png,I0445_01_06.png,2.971684612 1404 | I0445_01_01.png,I0445_01_05.png,3.34824407 1405 | I0445_01_02.png,I0445_01_05.png,7.20945791 1406 | I0445_01_01.png,I0445_01_03.png,8.308734596 1407 | I0445_01_02.png,I0445_01_03.png,6.59651079 1408 | I0445_01_02.png,I0445_01_06.png,3.589200111 1409 | I0445_01_03.png,I0445_01_06.png,7.114161251 1410 | I0445_01_01.png,I0445_01_02.png,8.721086014 1411 | I0448_01_03.png,I0448_01_05.png,2.418724585 1412 | I0448_01_02.png,I0448_01_06.png,3.81395849 1413 | I0448_01_03.png,I0448_01_04.png,4.556861353 1414 | I0448_01_02.png,I0448_01_04.png,4.539499325 1415 | I0448_01_01.png,I0448_01_06.png,6.812225093 1416 | I0448_01_01.png,I0448_01_04.png,7.820748794 1417 | I0448_01_02.png,I0448_01_03.png,5.67349126 1418 | I0448_01_01.png,I0448_01_05.png,4.684149558 1419 | I0448_01_05.png,I0448_01_06.png,3.085668403 1420 | I0448_01_04.png,I0448_01_05.png,4.47865303 1421 | I0448_01_04.png,I0448_01_06.png,3.580800938 1422 | I0448_01_01.png,I0448_01_02.png,7.118775388 1423 | I0448_01_01.png,I0448_01_03.png,4.074208434 1424 | I0448_01_02.png,I0448_01_05.png,4.569951764 1425 | I0448_01_03.png,I0448_01_06.png,3.406376555 1426 | I0456_01_01.png,I0456_01_04.png,3.410337401 1427 | I0456_01_05.png,I0456_01_06.png,6.968920651 1428 | I0456_01_02.png,I0456_01_04.png,2.551156753 1429 | I0456_01_04.png,I0456_01_06.png,4.507019682 1430 | I0456_01_01.png,I0456_01_05.png,4.529031106 1431 | I0456_01_02.png,I0456_01_06.png,4.657652618 1432 | I0456_01_01.png,I0456_01_03.png,4.196871415 1433 | I0456_01_02.png,I0456_01_05.png,4.420540947 1434 | I0456_01_03.png,I0456_01_04.png,3.425507993 1435 | I0456_01_01.png,I0456_01_02.png,3.979851431 1436 | I0456_01_03.png,I0456_01_05.png,4.381478151 1437 | I0456_01_02.png,I0456_01_03.png,4.442618797 1438 | I0456_01_04.png,I0456_01_05.png,2.622285269 1439 | I0456_01_01.png,I0456_01_06.png,3.042226646 1440 | I0456_01_03.png,I0456_01_06.png,4.507197634 1441 | I0459_01_01.png,I0459_01_05.png,2.059099671 1442 | I0459_01_02.png,I0459_01_04.png,2.684414146 1443 | I0459_01_02.png,I0459_01_05.png,2.239076927 1444 | I0459_01_04.png,I0459_01_05.png,2.310841375 1445 | I0459_01_01.png,I0459_01_06.png,2.5695504 1446 | I0459_01_02.png,I0459_01_06.png,2.209336024 1447 | I0459_01_02.png,I0459_01_03.png,3.643737691 1448 | I0459_01_01.png,I0459_01_03.png,2.473798462 1449 | I0459_01_03.png,I0459_01_04.png,3.057993092 1450 | I0459_01_01.png,I0459_01_04.png,2.620229696 1451 | I0459_01_01.png,I0459_01_02.png,2.303698317 1452 | I0459_01_05.png,I0459_01_06.png,2.714246275 1453 | I0459_01_03.png,I0459_01_06.png,3.729458638 1454 | I0459_01_03.png,I0459_01_05.png,2.960326896 1455 | I0459_01_04.png,I0459_01_06.png,3.109899689 1456 | I0462_01_02.png,I0462_01_05.png,6.309619211 1457 | I0462_01_03.png,I0462_01_04.png,3.256516866 1458 | I0462_01_01.png,I0462_01_05.png,4.347887807 1459 | I0462_01_05.png,I0462_01_06.png,4.945166549 1460 | I0462_01_02.png,I0462_01_04.png,2.370363219 1461 | I0462_01_02.png,I0462_01_03.png,3.258652895 1462 | I0462_01_01.png,I0462_01_03.png,2.800149652 1463 | I0462_01_04.png,I0462_01_05.png,5.24379443 1464 | I0462_01_02.png,I0462_01_06.png,2.34017122 1465 | I0462_01_01.png,I0462_01_02.png,3.769088428 1466 | I0462_01_01.png,I0462_01_06.png,3.067851505 1467 | I0462_01_03.png,I0462_01_06.png,4.41461054 1468 | I0462_01_01.png,I0462_01_04.png,2.795332413 1469 | I0462_01_03.png,I0462_01_05.png,5.159282229 1470 | I0462_01_04.png,I0462_01_06.png,2.807500931 1471 | I0466_01_01.png,I0466_01_03.png,2.724876715 1472 | I0466_01_03.png,I0466_01_04.png,4.61501905 1473 | I0466_01_01.png,I0466_01_04.png,5.740442619 1474 | I0466_01_02.png,I0466_01_06.png,3.644697758 1475 | I0466_01_02.png,I0466_01_04.png,4.817566627 1476 | I0466_01_01.png,I0466_01_06.png,2.562516036 1477 | I0466_01_04.png,I0466_01_06.png,5.461463376 1478 | I0466_01_02.png,I0466_01_03.png,3.826038057 1479 | I0466_01_03.png,I0466_01_06.png,3.263953369 1480 | I0466_01_05.png,I0466_01_06.png,2.808531423 1481 | I0466_01_04.png,I0466_01_05.png,4.788450793 1482 | I0466_01_01.png,I0466_01_05.png,2.746590516 1483 | I0466_01_01.png,I0466_01_02.png,3.256752216 1484 | I0466_01_02.png,I0466_01_05.png,3.377049688 1485 | I0466_01_03.png,I0466_01_05.png,2.725389577 1486 | I0467_01_02.png,I0467_01_03.png,6.464846073 1487 | I0467_01_01.png,I0467_01_05.png,3.932929068 1488 | I0467_01_02.png,I0467_01_06.png,3.277055767 1489 | I0467_01_05.png,I0467_01_06.png,3.852815058 1490 | I0467_01_04.png,I0467_01_05.png,4.373031202 1491 | I0467_01_04.png,I0467_01_06.png,2.338914725 1492 | I0467_01_01.png,I0467_01_06.png,5.486143387 1493 | I0467_01_01.png,I0467_01_02.png,6.659475075 1494 | I0467_01_03.png,I0467_01_04.png,4.241970334 1495 | I0467_01_02.png,I0467_01_04.png,3.651403396 1496 | I0467_01_02.png,I0467_01_05.png,4.84123983 1497 | I0467_01_01.png,I0467_01_04.png,4.295415904 1498 | I0467_01_01.png,I0467_01_03.png,5.63076474 1499 | I0467_01_03.png,I0467_01_05.png,6.143156611 1500 | I0467_01_03.png,I0467_01_06.png,4.303713493 1501 | I0471_01_01.png,I0471_01_06.png,2.419949684 1502 | I0471_01_04.png,I0471_01_06.png,2.646396696 1503 | I0471_01_02.png,I0471_01_04.png,3.513297848 1504 | I0471_01_02.png,I0471_01_03.png,2.309422829 1505 | I0471_01_05.png,I0471_01_06.png,2.820495302 1506 | I0471_01_04.png,I0471_01_05.png,2.970307941 1507 | I0471_01_03.png,I0471_01_04.png,2.458392724 1508 | I0471_01_01.png,I0471_01_03.png,2.501277147 1509 | I0471_01_01.png,I0471_01_05.png,2.660552381 1510 | I0471_01_02.png,I0471_01_05.png,2.514642399 1511 | I0471_01_03.png,I0471_01_05.png,3.126738063 1512 | I0471_01_01.png,I0471_01_02.png,2.092126099 1513 | I0471_01_03.png,I0471_01_06.png,2.522087809 1514 | I0471_01_02.png,I0471_01_06.png,2.590747472 1515 | I0471_01_01.png,I0471_01_04.png,2.700832929 1516 | I0473_01_01.png,I0473_01_04.png,3.156235641 1517 | I0473_01_02.png,I0473_01_06.png,3.660405823 1518 | I0473_01_01.png,I0473_01_06.png,4.951133859 1519 | I0473_01_04.png,I0473_01_06.png,2.694832485 1520 | I0473_01_02.png,I0473_01_04.png,2.943413071 1521 | I0473_01_03.png,I0473_01_06.png,5.157912211 1522 | I0473_01_01.png,I0473_01_03.png,2.730626749 1523 | I0473_01_01.png,I0473_01_05.png,3.2035074 1524 | I0473_01_03.png,I0473_01_04.png,3.901841015 1525 | I0473_01_05.png,I0473_01_06.png,7.502484776 1526 | I0473_01_04.png,I0473_01_05.png,5.380542375 1527 | I0473_01_01.png,I0473_01_02.png,4.206142566 1528 | I0473_01_03.png,I0473_01_05.png,3.612690259 1529 | I0473_01_02.png,I0473_01_05.png,6.54918081 1530 | I0473_01_02.png,I0473_01_03.png,4.567302158 1531 | I0486_01_03.png,I0486_01_05.png,5.120815252 1532 | I0486_01_04.png,I0486_01_06.png,4.040892433 1533 | I0486_01_01.png,I0486_01_03.png,5.254458651 1534 | I0486_01_01.png,I0486_01_02.png,2.497607253 1535 | I0486_01_03.png,I0486_01_06.png,4.507851553 1536 | I0486_01_01.png,I0486_01_05.png,3.097132732 1537 | I0486_01_03.png,I0486_01_04.png,5.152575282 1538 | I0486_01_02.png,I0486_01_03.png,5.424791056 1539 | I0486_01_01.png,I0486_01_04.png,3.454439824 1540 | I0486_01_04.png,I0486_01_05.png,3.913095656 1541 | I0486_01_02.png,I0486_01_04.png,3.726028375 1542 | I0486_01_05.png,I0486_01_06.png,3.182636105 1543 | I0486_01_02.png,I0486_01_05.png,2.363038683 1544 | I0486_01_02.png,I0486_01_06.png,2.987635153 1545 | I0486_01_01.png,I0486_01_06.png,3.797597725 1546 | I0487_01_01.png,I0487_01_02.png,3.849762418 1547 | I0487_01_01.png,I0487_01_06.png,4.033496267 1548 | I0487_01_01.png,I0487_01_04.png,3.529086752 1549 | I0487_01_02.png,I0487_01_03.png,5.923992062 1550 | I0487_01_02.png,I0487_01_05.png,4.878943697 1551 | I0487_01_03.png,I0487_01_06.png,3.612566732 1552 | I0487_01_04.png,I0487_01_05.png,4.877365037 1553 | I0487_01_03.png,I0487_01_04.png,4.76495291 1554 | I0487_01_02.png,I0487_01_04.png,3.725787528 1555 | I0487_01_01.png,I0487_01_05.png,4.075470013 1556 | I0487_01_05.png,I0487_01_06.png,3.115688192 1557 | I0487_01_04.png,I0487_01_06.png,3.975414634 1558 | I0487_01_02.png,I0487_01_06.png,4.824687895 1559 | I0487_01_03.png,I0487_01_05.png,3.464749586 1560 | I0487_01_01.png,I0487_01_03.png,5.090240082 1561 | I0489_01_05.png,I0489_01_06.png,3.503198543 1562 | I0489_01_01.png,I0489_01_04.png,4.197504826 1563 | I0489_01_02.png,I0489_01_03.png,4.199590321 1564 | I0489_01_01.png,I0489_01_02.png,3.819527649 1565 | I0489_01_01.png,I0489_01_06.png,6.534632693 1566 | I0489_01_03.png,I0489_01_05.png,3.164921022 1567 | I0489_01_03.png,I0489_01_04.png,3.160135143 1568 | I0489_01_02.png,I0489_01_04.png,6.1337355 1569 | I0489_01_01.png,I0489_01_03.png,2.563143842 1570 | I0489_01_02.png,I0489_01_06.png,2.670667466 1571 | I0489_01_02.png,I0489_01_05.png,2.408527144 1572 | I0489_01_01.png,I0489_01_05.png,3.527595181 1573 | I0489_01_04.png,I0489_01_06.png,8.057103173 1574 | I0489_01_03.png,I0489_01_06.png,5.916957674 1575 | I0489_01_04.png,I0489_01_05.png,5.396162178 1576 | I0496_01_02.png,I0496_01_06.png,3.521815491 1577 | I0496_01_04.png,I0496_01_06.png,5.15095871 1578 | I0496_01_02.png,I0496_01_03.png,2.735431386 1579 | I0496_01_03.png,I0496_01_05.png,3.465068246 1580 | I0496_01_04.png,I0496_01_05.png,5.900184337 1581 | I0496_01_01.png,I0496_01_06.png,2.951834371 1582 | I0496_01_01.png,I0496_01_05.png,2.780770115 1583 | I0496_01_01.png,I0496_01_04.png,6.968531681 1584 | I0496_01_01.png,I0496_01_03.png,2.646580537 1585 | I0496_01_03.png,I0496_01_04.png,7.053594331 1586 | I0496_01_03.png,I0496_01_06.png,2.621069485 1587 | I0496_01_05.png,I0496_01_06.png,4.218315666 1588 | I0496_01_02.png,I0496_01_04.png,7.988241546 1589 | I0496_01_01.png,I0496_01_02.png,2.649400915 1590 | I0496_01_02.png,I0496_01_05.png,2.598946575 1591 | I0506_01_03.png,I0506_01_05.png,5.428730359 1592 | I0506_01_01.png,I0506_01_05.png,6.097642111 1593 | I0506_01_05.png,I0506_01_06.png,4.899935084 1594 | I0506_01_02.png,I0506_01_04.png,3.6042424 1595 | I0506_01_03.png,I0506_01_06.png,7.32075864 1596 | I0506_01_04.png,I0506_01_06.png,4.444040604 1597 | I0506_01_01.png,I0506_01_03.png,8.179438717 1598 | I0506_01_01.png,I0506_01_02.png,2.446775319 1599 | I0506_01_01.png,I0506_01_06.png,4.092689344 1600 | I0506_01_02.png,I0506_01_06.png,3.758928965 1601 | I0506_01_03.png,I0506_01_04.png,7.785678069 1602 | I0506_01_02.png,I0506_01_05.png,5.694341071 1603 | I0506_01_04.png,I0506_01_05.png,5.690944255 1604 | I0506_01_01.png,I0506_01_04.png,4.539672142 1605 | I0506_01_02.png,I0506_01_03.png,7.919832196 1606 | I0520_01_02.png,I0520_01_06.png,4.464632903 1607 | I0520_01_03.png,I0520_01_05.png,5.59854509 1608 | I0520_01_04.png,I0520_01_06.png,3.502809444 1609 | I0520_01_01.png,I0520_01_06.png,4.175266152 1610 | I0520_01_04.png,I0520_01_05.png,3.192325744 1611 | I0520_01_01.png,I0520_01_02.png,5.6488988 1612 | I0520_01_05.png,I0520_01_06.png,4.786888023 1613 | I0520_01_03.png,I0520_01_04.png,4.905542189 1614 | I0520_01_02.png,I0520_01_04.png,3.299892926 1615 | I0520_01_01.png,I0520_01_03.png,7.352575412 1616 | I0520_01_03.png,I0520_01_06.png,5.924539963 1617 | I0520_01_01.png,I0520_01_05.png,5.785720615 1618 | I0520_01_02.png,I0520_01_05.png,3.960131963 1619 | I0520_01_02.png,I0520_01_03.png,5.134779631 1620 | I0520_01_01.png,I0520_01_04.png,4.805144033 1621 | I0526_01_03.png,I0526_01_06.png,7.640635503 1622 | I0526_01_01.png,I0526_01_05.png,5.983442496 1623 | I0526_01_04.png,I0526_01_05.png,2.813394852 1624 | I0526_01_02.png,I0526_01_05.png,5.621121819 1625 | I0526_01_01.png,I0526_01_04.png,4.970847129 1626 | I0526_01_01.png,I0526_01_03.png,7.2416695 1627 | I0526_01_03.png,I0526_01_04.png,4.082890288 1628 | I0526_01_01.png,I0526_01_06.png,4.502126808 1629 | I0526_01_03.png,I0526_01_05.png,2.556720992 1630 | I0526_01_02.png,I0526_01_06.png,4.474797693 1631 | I0526_01_04.png,I0526_01_06.png,5.512694131 1632 | I0526_01_01.png,I0526_01_02.png,5.426545572 1633 | I0526_01_02.png,I0526_01_03.png,7.308175598 1634 | I0526_01_02.png,I0526_01_04.png,6.99765678 1635 | I0526_01_05.png,I0526_01_06.png,6.182359005 1636 | I0537_01_04.png,I0537_01_05.png,4.604769438 1637 | I0537_01_01.png,I0537_01_02.png,6.013718638 1638 | I0537_01_03.png,I0537_01_05.png,5.341894796 1639 | I0537_01_03.png,I0537_01_04.png,6.827126506 1640 | I0537_01_02.png,I0537_01_04.png,3.87069967 1641 | I0537_01_04.png,I0537_01_06.png,5.426754731 1642 | I0537_01_01.png,I0537_01_04.png,6.451669332 1643 | I0537_01_03.png,I0537_01_06.png,5.157579114 1644 | I0537_01_02.png,I0537_01_06.png,5.644319326 1645 | I0537_01_05.png,I0537_01_06.png,4.507711237 1646 | I0537_01_01.png,I0537_01_05.png,3.498525443 1647 | I0537_01_01.png,I0537_01_06.png,4.957901554 1648 | I0537_01_02.png,I0537_01_05.png,4.979300136 1649 | I0537_01_01.png,I0537_01_03.png,5.390811379 1650 | I0537_01_02.png,I0537_01_03.png,7.124993898 1651 | I0538_01_01.png,I0538_01_03.png,5.234062171 1652 | I0538_01_02.png,I0538_01_06.png,6.683739414 1653 | I0538_01_01.png,I0538_01_05.png,3.782871784 1654 | I0538_01_01.png,I0538_01_04.png,4.829081489 1655 | I0538_01_02.png,I0538_01_05.png,4.813465507 1656 | I0538_01_01.png,I0538_01_02.png,5.825421198 1657 | I0538_01_05.png,I0538_01_06.png,3.423027114 1658 | I0538_01_03.png,I0538_01_04.png,4.731228843 1659 | I0538_01_03.png,I0538_01_05.png,2.963489346 1660 | I0538_01_02.png,I0538_01_03.png,6.66507415 1661 | I0538_01_02.png,I0538_01_04.png,4.155338014 1662 | I0538_01_03.png,I0538_01_06.png,3.643759149 1663 | I0538_01_04.png,I0538_01_06.png,3.696616307 1664 | I0538_01_01.png,I0538_01_06.png,4.868739687 1665 | I0538_01_04.png,I0538_01_05.png,3.572453192 1666 | I0541_01_01.png,I0541_01_05.png,3.410965045 1667 | I0541_01_02.png,I0541_01_05.png,3.566906669 1668 | I0541_01_03.png,I0541_01_06.png,3.079232675 1669 | I0541_01_02.png,I0541_01_03.png,2.682961603 1670 | I0541_01_01.png,I0541_01_06.png,2.959328472 1671 | I0541_01_03.png,I0541_01_05.png,3.439960517 1672 | I0541_01_04.png,I0541_01_05.png,5.991747835 1673 | I0541_01_03.png,I0541_01_04.png,8.216664863 1674 | I0541_01_01.png,I0541_01_04.png,6.232810206 1675 | I0541_01_02.png,I0541_01_04.png,7.238527629 1676 | I0541_01_05.png,I0541_01_06.png,2.439377762 1677 | I0541_01_02.png,I0541_01_06.png,3.165611384 1678 | I0541_01_01.png,I0541_01_02.png,3.634295961 1679 | I0541_01_04.png,I0541_01_06.png,6.809078997 1680 | I0541_01_01.png,I0541_01_03.png,3.095619279 1681 | I0546_01_01.png,I0546_01_06.png,5.169969081 1682 | I0546_01_04.png,I0546_01_06.png,8.306261137 1683 | I0546_01_02.png,I0546_01_03.png,6.536100082 1684 | I0546_01_02.png,I0546_01_04.png,9.895108065 1685 | I0546_01_03.png,I0546_01_04.png,8.791524574 1686 | I0546_01_01.png,I0546_01_03.png,3.569263368 1687 | I0546_01_02.png,I0546_01_06.png,6.223159426 1688 | I0546_01_01.png,I0546_01_02.png,6.896372121 1689 | I0546_01_02.png,I0546_01_05.png,6.360081342 1690 | I0546_01_03.png,I0546_01_05.png,3.069298061 1691 | I0546_01_04.png,I0546_01_05.png,8.047034753 1692 | I0546_01_03.png,I0546_01_06.png,5.4009272 1693 | I0546_01_01.png,I0546_01_05.png,2.929510404 1694 | I0546_01_05.png,I0546_01_06.png,3.592588416 1695 | I0546_01_01.png,I0546_01_04.png,7.74564212 1696 | I0550_01_05.png,I0550_01_06.png,3.335393433 1697 | I0550_01_03.png,I0550_01_05.png,2.87215553 1698 | I0550_01_01.png,I0550_01_03.png,4.79589921 1699 | I0550_01_01.png,I0550_01_04.png,3.585600224 1700 | I0550_01_03.png,I0550_01_06.png,3.356114397 1701 | I0550_01_02.png,I0550_01_06.png,3.559656506 1702 | I0550_01_03.png,I0550_01_04.png,6.087340525 1703 | I0550_01_02.png,I0550_01_04.png,3.776311938 1704 | I0550_01_04.png,I0550_01_05.png,5.283683567 1705 | I0550_01_01.png,I0550_01_05.png,3.567619566 1706 | I0550_01_02.png,I0550_01_05.png,3.517057805 1707 | I0550_01_01.png,I0550_01_02.png,5.016636735 1708 | I0550_01_04.png,I0550_01_06.png,5.631050839 1709 | I0550_01_01.png,I0550_01_06.png,6.003922874 1710 | I0550_01_02.png,I0550_01_03.png,4.039765925 1711 | I0556_01_05.png,I0556_01_06.png,3.018558397 1712 | I0556_01_01.png,I0556_01_06.png,3.733568194 1713 | I0556_01_01.png,I0556_01_02.png,7.107432491 1714 | I0556_01_04.png,I0556_01_06.png,5.651916887 1715 | I0556_01_03.png,I0556_01_04.png,3.968661793 1716 | I0556_01_01.png,I0556_01_05.png,2.512411752 1717 | I0556_01_01.png,I0556_01_04.png,7.298263702 1718 | I0556_01_02.png,I0556_01_05.png,6.37478682 1719 | I0556_01_01.png,I0556_01_03.png,5.573955527 1720 | I0556_01_04.png,I0556_01_05.png,6.046351705 1721 | I0556_01_02.png,I0556_01_06.png,8.42184046 1722 | I0556_01_03.png,I0556_01_06.png,3.944549477 1723 | I0556_01_03.png,I0556_01_05.png,3.54144364 1724 | I0556_01_02.png,I0556_01_03.png,10.60719006 1725 | I0556_01_02.png,I0556_01_04.png,12.50416435 1726 | I0559_01_02.png,I0559_01_05.png,4.235786728 1727 | I0559_01_01.png,I0559_01_04.png,4.209666296 1728 | I0559_01_02.png,I0559_01_03.png,5.466442281 1729 | I0559_01_03.png,I0559_01_04.png,7.09324185 1730 | I0559_01_03.png,I0559_01_05.png,5.875864811 1731 | I0559_01_02.png,I0559_01_06.png,4.54216092 1732 | I0559_01_03.png,I0559_01_06.png,5.687496996 1733 | I0559_01_01.png,I0559_01_05.png,6.147884572 1734 | I0559_01_01.png,I0559_01_06.png,7.067193225 1735 | I0559_01_05.png,I0559_01_06.png,3.220747159 1736 | I0559_01_04.png,I0559_01_05.png,5.146930059 1737 | I0559_01_01.png,I0559_01_03.png,4.868729877 1738 | I0559_01_04.png,I0559_01_06.png,4.359306828 1739 | I0559_01_02.png,I0559_01_04.png,4.517239375 1740 | I0559_01_01.png,I0559_01_02.png,4.251035781 1741 | I0561_01_02.png,I0561_01_03.png,7.496413628 1742 | I0561_01_03.png,I0561_01_04.png,5.118145771 1743 | I0561_01_03.png,I0561_01_05.png,4.803518112 1744 | I0561_01_01.png,I0561_01_04.png,4.653315305 1745 | I0561_01_03.png,I0561_01_06.png,4.924675092 1746 | I0561_01_02.png,I0561_01_06.png,6.745704855 1747 | I0561_01_01.png,I0561_01_02.png,9.247293496 1748 | I0561_01_01.png,I0561_01_06.png,5.607998023 1749 | I0561_01_02.png,I0561_01_04.png,9.344549559 1750 | I0561_01_04.png,I0561_01_05.png,5.664091537 1751 | I0561_01_01.png,I0561_01_03.png,4.523325201 1752 | I0561_01_02.png,I0561_01_05.png,7.663413299 1753 | I0561_01_04.png,I0561_01_06.png,5.451823666 1754 | I0561_01_01.png,I0561_01_05.png,4.138063456 1755 | I0561_01_05.png,I0561_01_06.png,4.310828131 1756 | I0565_01_01.png,I0565_01_02.png,7.624385705 1757 | I0565_01_03.png,I0565_01_06.png,3.028190976 1758 | I0565_01_02.png,I0565_01_04.png,8.883601179 1759 | I0565_01_04.png,I0565_01_05.png,4.912675118 1760 | I0565_01_03.png,I0565_01_05.png,3.022624975 1761 | I0565_01_01.png,I0565_01_04.png,3.755026736 1762 | I0565_01_02.png,I0565_01_03.png,7.673046609 1763 | I0565_01_02.png,I0565_01_05.png,6.688686241 1764 | I0565_01_01.png,I0565_01_03.png,3.744670906 1765 | I0565_01_01.png,I0565_01_05.png,3.8036723 1766 | I0565_01_05.png,I0565_01_06.png,2.630205297 1767 | I0565_01_03.png,I0565_01_04.png,4.706377366 1768 | I0565_01_02.png,I0565_01_06.png,5.130875367 1769 | I0565_01_01.png,I0565_01_06.png,3.63678862 1770 | I0565_01_04.png,I0565_01_06.png,6.492334159 1771 | I0568_01_05.png,I0568_01_06.png,3.687054639 1772 | I0568_01_01.png,I0568_01_05.png,2.983313896 1773 | I0568_01_01.png,I0568_01_03.png,5.342655345 1774 | I0568_01_02.png,I0568_01_04.png,4.010759688 1775 | I0568_01_02.png,I0568_01_06.png,3.067584631 1776 | I0568_01_01.png,I0568_01_04.png,2.701121616 1777 | I0568_01_03.png,I0568_01_06.png,3.572010798 1778 | I0568_01_01.png,I0568_01_06.png,3.819850299 1779 | I0568_01_02.png,I0568_01_03.png,4.149511579 1780 | I0568_01_03.png,I0568_01_04.png,3.343217382 1781 | I0568_01_02.png,I0568_01_05.png,3.539242417 1782 | I0568_01_01.png,I0568_01_02.png,3.876017349 1783 | I0568_01_03.png,I0568_01_05.png,3.959059008 1784 | I0568_01_04.png,I0568_01_06.png,3.157001607 1785 | I0568_01_04.png,I0568_01_05.png,3.184775187 1786 | I0569_01_02.png,I0569_01_06.png,7.447211816 1787 | I0569_01_01.png,I0569_01_04.png,3.152405536 1788 | I0569_01_03.png,I0569_01_06.png,6.19196063 1789 | I0569_01_04.png,I0569_01_06.png,3.215010237 1790 | I0569_01_03.png,I0569_01_05.png,3.256282532 1791 | I0569_01_04.png,I0569_01_05.png,3.08962219 1792 | I0569_01_02.png,I0569_01_03.png,5.108399138 1793 | I0569_01_05.png,I0569_01_06.png,5.351309095 1794 | I0569_01_02.png,I0569_01_04.png,5.208901276 1795 | I0569_01_02.png,I0569_01_05.png,3.936872845 1796 | I0569_01_01.png,I0569_01_03.png,3.772100087 1797 | I0569_01_01.png,I0569_01_02.png,6.489680407 1798 | I0569_01_03.png,I0569_01_04.png,3.563833325 1799 | I0569_01_01.png,I0569_01_06.png,3.377912238 1800 | I0569_01_01.png,I0569_01_05.png,3.84949768 1801 | I0580_01_02.png,I0580_01_03.png,3.194010038 1802 | I0580_01_02.png,I0580_01_06.png,2.506468803 1803 | I0580_01_04.png,I0580_01_05.png,3.290705077 1804 | I0580_01_02.png,I0580_01_04.png,2.507012929 1805 | I0580_01_01.png,I0580_01_03.png,4.062720893 1806 | I0580_01_01.png,I0580_01_05.png,3.637559666 1807 | I0580_01_03.png,I0580_01_05.png,2.808642825 1808 | I0580_01_03.png,I0580_01_06.png,3.23213297 1809 | I0580_01_01.png,I0580_01_06.png,3.664681149 1810 | I0580_01_01.png,I0580_01_02.png,4.397845836 1811 | I0580_01_02.png,I0580_01_05.png,3.035990507 1812 | I0580_01_03.png,I0580_01_04.png,3.290274177 1813 | I0580_01_05.png,I0580_01_06.png,3.461250641 1814 | I0580_01_01.png,I0580_01_04.png,4.174746217 1815 | I0580_01_04.png,I0580_01_06.png,2.147490182 1816 | I0582_01_02.png,I0582_01_04.png,7.647613099 1817 | I0582_01_01.png,I0582_01_03.png,4.147248366 1818 | I0582_01_01.png,I0582_01_02.png,5.548827433 1819 | I0582_01_04.png,I0582_01_06.png,6.65458425 1820 | I0582_01_03.png,I0582_01_04.png,7.224242664 1821 | I0582_01_03.png,I0582_01_05.png,3.762462098 1822 | I0582_01_01.png,I0582_01_05.png,5.331618097 1823 | I0582_01_01.png,I0582_01_04.png,5.692047803 1824 | I0582_01_02.png,I0582_01_03.png,4.886366684 1825 | I0582_01_02.png,I0582_01_05.png,5.323692815 1826 | I0582_01_04.png,I0582_01_05.png,7.782055923 1827 | I0582_01_03.png,I0582_01_06.png,2.448438652 1828 | I0582_01_02.png,I0582_01_06.png,3.568760998 1829 | I0582_01_05.png,I0582_01_06.png,4.769623849 1830 | I0582_01_01.png,I0582_01_06.png,3.880424247 1831 | I0585_01_03.png,I0585_01_05.png,3.582005559 1832 | I0585_01_02.png,I0585_01_06.png,3.649655369 1833 | I0585_01_03.png,I0585_01_04.png,4.729963436 1834 | I0585_01_01.png,I0585_01_03.png,2.916417709 1835 | I0585_01_01.png,I0585_01_05.png,4.14148038 1836 | I0585_01_01.png,I0585_01_02.png,4.69844561 1837 | I0585_01_02.png,I0585_01_05.png,3.119238391 1838 | I0585_01_04.png,I0585_01_05.png,5.663050746 1839 | I0585_01_03.png,I0585_01_06.png,3.340245709 1840 | I0585_01_02.png,I0585_01_03.png,4.971470424 1841 | I0585_01_01.png,I0585_01_04.png,4.623519585 1842 | I0585_01_02.png,I0585_01_04.png,6.862120894 1843 | I0585_01_04.png,I0585_01_06.png,6.074927069 1844 | I0585_01_01.png,I0585_01_06.png,4.303134345 1845 | I0585_01_05.png,I0585_01_06.png,2.078504033 1846 | I0605_01_01.png,I0605_01_04.png,3.424159475 1847 | I0605_01_02.png,I0605_01_06.png,3.290955351 1848 | I0605_01_02.png,I0605_01_04.png,3.662720484 1849 | I0605_01_01.png,I0605_01_05.png,2.704210369 1850 | I0605_01_04.png,I0605_01_05.png,3.066508384 1851 | I0605_01_02.png,I0605_01_05.png,3.225148223 1852 | I0605_01_02.png,I0605_01_03.png,2.993933743 1853 | I0605_01_01.png,I0605_01_03.png,3.144403645 1854 | I0605_01_01.png,I0605_01_06.png,3.298615176 1855 | I0605_01_03.png,I0605_01_04.png,3.695920548 1856 | I0605_01_05.png,I0605_01_06.png,2.633239049 1857 | I0605_01_03.png,I0605_01_06.png,3.217637436 1858 | I0605_01_04.png,I0605_01_06.png,2.518535965 1859 | I0605_01_03.png,I0605_01_05.png,3.500436492 1860 | I0605_01_01.png,I0605_01_02.png,2.591863899 1861 | I0606_01_01.png,I0606_01_04.png,6.592434755 1862 | I0606_01_01.png,I0606_01_05.png,4.331651038 1863 | I0606_01_02.png,I0606_01_03.png,3.963880376 1864 | I0606_01_01.png,I0606_01_03.png,4.265204165 1865 | I0606_01_02.png,I0606_01_04.png,6.488830573 1866 | I0606_01_01.png,I0606_01_06.png,4.024399772 1867 | I0606_01_03.png,I0606_01_05.png,4.333974796 1868 | I0606_01_03.png,I0606_01_04.png,6.089396579 1869 | I0606_01_02.png,I0606_01_06.png,3.19230052 1870 | I0606_01_05.png,I0606_01_06.png,3.579510759 1871 | I0606_01_04.png,I0606_01_05.png,4.771285977 1872 | I0606_01_03.png,I0606_01_06.png,4.138481553 1873 | I0606_01_02.png,I0606_01_05.png,3.805181444 1874 | I0606_01_04.png,I0606_01_06.png,6.874012343 1875 | I0606_01_01.png,I0606_01_02.png,3.275969441 1876 | I0610_01_03.png,I0610_01_05.png,2.870483965 1877 | I0610_01_01.png,I0610_01_05.png,3.728014893 1878 | I0610_01_02.png,I0610_01_04.png,3.784393717 1879 | I0610_01_02.png,I0610_01_03.png,4.462749917 1880 | I0610_01_01.png,I0610_01_02.png,5.467272205 1881 | I0610_01_01.png,I0610_01_04.png,6.205998161 1882 | I0610_01_03.png,I0610_01_04.png,5.202757109 1883 | I0610_01_01.png,I0610_01_03.png,3.641242064 1884 | I0610_01_04.png,I0610_01_06.png,4.872437862 1885 | I0610_01_03.png,I0610_01_06.png,3.952198751 1886 | I0610_01_01.png,I0610_01_06.png,5.539252793 1887 | I0610_01_04.png,I0610_01_05.png,5.230782135 1888 | I0610_01_05.png,I0610_01_06.png,2.96512569 1889 | I0610_01_02.png,I0610_01_05.png,3.003736175 1890 | I0610_01_02.png,I0610_01_06.png,2.673143129 1891 | I0618_01_01.png,I0618_01_06.png,4.373509727 1892 | I0618_01_04.png,I0618_01_05.png,2.46801501 1893 | I0618_01_01.png,I0618_01_05.png,5.109287418 1894 | I0618_01_02.png,I0618_01_06.png,9.377928037 1895 | I0618_01_03.png,I0618_01_04.png,4.181295906 1896 | I0618_01_02.png,I0618_01_04.png,8.30375243 1897 | I0618_01_05.png,I0618_01_06.png,4.825090142 1898 | I0618_01_01.png,I0618_01_03.png,4.129604071 1899 | I0618_01_03.png,I0618_01_06.png,3.559636991 1900 | I0618_01_02.png,I0618_01_05.png,8.303823013 1901 | I0618_01_01.png,I0618_01_02.png,9.547360547 1902 | I0618_01_04.png,I0618_01_06.png,4.623341872 1903 | I0618_01_02.png,I0618_01_03.png,7.905404839 1904 | I0618_01_03.png,I0618_01_05.png,4.064174146 1905 | I0618_01_01.png,I0618_01_04.png,5.226636352 1906 | I0625_01_02.png,I0625_01_04.png,2.944828877 1907 | I0625_01_01.png,I0625_01_05.png,3.265604132 1908 | I0625_01_05.png,I0625_01_06.png,3.684951109 1909 | I0625_01_02.png,I0625_01_05.png,3.590291441 1910 | I0625_01_01.png,I0625_01_02.png,3.888295677 1911 | I0625_01_02.png,I0625_01_03.png,5.61494349 1912 | I0625_01_04.png,I0625_01_06.png,4.686686949 1913 | I0625_01_02.png,I0625_01_06.png,3.172387312 1914 | I0625_01_03.png,I0625_01_05.png,4.714640592 1915 | I0625_01_01.png,I0625_01_03.png,6.579555458 1916 | I0625_01_01.png,I0625_01_04.png,3.228168618 1917 | I0625_01_04.png,I0625_01_05.png,2.842596629 1918 | I0625_01_03.png,I0625_01_06.png,4.524574948 1919 | I0625_01_03.png,I0625_01_04.png,4.951016106 1920 | I0625_01_01.png,I0625_01_06.png,2.994086809 1921 | I0634_01_01.png,I0634_01_02.png,5.704177471 1922 | I0634_01_02.png,I0634_01_05.png,4.876759318 1923 | I0634_01_01.png,I0634_01_05.png,4.015449587 1924 | I0634_01_04.png,I0634_01_06.png,5.681332591 1925 | I0634_01_03.png,I0634_01_04.png,3.29165402 1926 | I0634_01_02.png,I0634_01_04.png,7.125929577 1927 | I0634_01_01.png,I0634_01_03.png,3.158321375 1928 | I0634_01_05.png,I0634_01_06.png,4.20731512 1929 | I0634_01_01.png,I0634_01_06.png,4.12576879 1930 | I0634_01_03.png,I0634_01_06.png,3.194415983 1931 | I0634_01_04.png,I0634_01_05.png,4.089259379 1932 | I0634_01_03.png,I0634_01_05.png,4.404238356 1933 | I0634_01_02.png,I0634_01_03.png,5.196845917 1934 | I0634_01_02.png,I0634_01_06.png,3.19944499 1935 | I0634_01_01.png,I0634_01_04.png,4.431375235 1936 | I0642_01_04.png,I0642_01_05.png,4.033768907 1937 | I0642_01_01.png,I0642_01_04.png,5.263027551 1938 | I0642_01_01.png,I0642_01_03.png,6.751888331 1939 | I0642_01_04.png,I0642_01_06.png,5.471816892 1940 | I0642_01_02.png,I0642_01_06.png,4.352517199 1941 | I0642_01_03.png,I0642_01_04.png,4.55715131 1942 | I0642_01_05.png,I0642_01_06.png,4.980022662 1943 | I0642_01_03.png,I0642_01_05.png,4.60088774 1944 | I0642_01_01.png,I0642_01_02.png,3.546726206 1945 | I0642_01_01.png,I0642_01_05.png,3.879097064 1946 | I0642_01_02.png,I0642_01_03.png,6.510842999 1947 | I0642_01_03.png,I0642_01_06.png,6.554804646 1948 | I0642_01_01.png,I0642_01_06.png,5.79196837 1949 | I0642_01_02.png,I0642_01_05.png,4.713964523 1950 | I0642_01_02.png,I0642_01_04.png,5.200226789 1951 | I0655_01_01.png,I0655_01_02.png,4.016917611 1952 | I0655_01_03.png,I0655_01_06.png,2.159310393 1953 | I0655_01_01.png,I0655_01_05.png,3.542096895 1954 | I0655_01_03.png,I0655_01_05.png,3.72852924 1955 | I0655_01_02.png,I0655_01_04.png,5.526144311 1956 | I0655_01_02.png,I0655_01_05.png,3.8496174 1957 | I0655_01_01.png,I0655_01_04.png,3.665557481 1958 | I0655_01_02.png,I0655_01_06.png,2.722604848 1959 | I0655_01_02.png,I0655_01_03.png,2.391659935 1960 | I0655_01_04.png,I0655_01_06.png,4.912534659 1961 | I0655_01_03.png,I0655_01_04.png,5.193610766 1962 | I0655_01_01.png,I0655_01_03.png,3.661309144 1963 | I0655_01_01.png,I0655_01_06.png,3.552003712 1964 | I0655_01_04.png,I0655_01_05.png,3.860834762 1965 | I0655_01_05.png,I0655_01_06.png,4.6611149 1966 | I0660_01_04.png,I0660_01_06.png,3.288560921 1967 | I0660_01_03.png,I0660_01_05.png,3.207346746 1968 | I0660_01_01.png,I0660_01_04.png,6.812192752 1969 | I0660_01_03.png,I0660_01_06.png,3.035244965 1970 | I0660_01_02.png,I0660_01_04.png,4.967660253 1971 | I0660_01_01.png,I0660_01_03.png,3.749779471 1972 | I0660_01_01.png,I0660_01_05.png,3.964651822 1973 | I0660_01_03.png,I0660_01_04.png,4.501927073 1974 | I0660_01_01.png,I0660_01_02.png,4.675516529 1975 | I0660_01_02.png,I0660_01_05.png,2.972488794 1976 | I0660_01_01.png,I0660_01_06.png,4.854012012 1977 | I0660_01_02.png,I0660_01_06.png,2.983366226 1978 | I0660_01_04.png,I0660_01_05.png,4.341914344 1979 | I0660_01_02.png,I0660_01_03.png,3.243319867 1980 | I0660_01_05.png,I0660_01_06.png,3.197142323 1981 | I0665_01_02.png,I0665_01_05.png,3.512037522 1982 | I0665_01_03.png,I0665_01_06.png,4.985530836 1983 | I0665_01_05.png,I0665_01_06.png,3.957398214 1984 | I0665_01_03.png,I0665_01_04.png,6.708469272 1985 | I0665_01_02.png,I0665_01_06.png,4.77462147 1986 | I0665_01_02.png,I0665_01_04.png,7.838596886 1987 | I0665_01_02.png,I0665_01_03.png,3.548969046 1988 | I0665_01_03.png,I0665_01_05.png,3.553202624 1989 | I0665_01_01.png,I0665_01_03.png,4.567526352 1990 | I0665_01_04.png,I0665_01_06.png,8.183040165 1991 | I0665_01_01.png,I0665_01_02.png,2.507282861 1992 | I0665_01_01.png,I0665_01_05.png,3.64144199 1993 | I0665_01_01.png,I0665_01_04.png,7.984479805 1994 | I0665_01_04.png,I0665_01_05.png,7.758788972 1995 | I0665_01_01.png,I0665_01_06.png,4.64112602 1996 | --------------------------------------------------------------------------------