├── .DS_Store
├── README.md
└── images
    ├── SC.png
    ├── multi_task_1.png
    ├── multi_task_2.png
    ├── t-SNE_a.png
    └── t-SNE_b.png


/.DS_Store:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Alpha-Innovator/SimChart9K/4afc2c5385d8bfd11f64e1356c25bffd20369114/.DS_Store


--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
 1 | [![arXiv](https://img.shields.io/badge/arXiv-2309.11268-b31b1b.svg)](https://arxiv.org/abs/2309.11268)
 2 | [![GitHub issues](https://img.shields.io/github/issues/Uni-Modal/SimChart9K)](https://github.com/Uni-Modal/SimChart9K/issues)
 3 | [![PRs Welcome](https://img.shields.io/badge/PRs-welcome-brightgreen.svg?style=flat-square)](https://github.com/Uni-Modal/SimChart9K/pulls)
 4 | 
 5 | # SimChart9K
 6 | 
 7 | <div align=center>
 8 | <img src="https://github.com/Uni-Modal/SimChart9K/blob/main/images/SC.png" height="100">
 9 | </div>
10 | 
11 | <div align="center">
12 | <h1>SimChart9K: An LLMs-based Simulatied Visual Chart Understanding Benchmark<br></h1>
13 | </div>
14 | 
15 | We perform data augmentation for chart perception and reasoning by leveraging an LLMs-based self-inspection data production scheme, producing the SimChart9K dataset, where the simulated dataset consists of 9,536 chart images and associated data annotations in CSV format. Besides, we observe that StructChart continuously improves the chart perception performance as more simulated charts are used for pre-training.
16 | 
17 | ## SimChart9K Dataset Download from google drive
18 | Downloading the official [SimChart9K](https://drive.google.com/file/d/1M_NA3sIJNwCUfqB1HH0p4lsnEm3NbvsI/view?usp=sharing) dataset from google drive
19 | 
20 | ## SimChart9K Dataset Download from Opendatalab
21 | 
22 | a. Register an account from OpenXLab website as follows.
23 | ```shell
24 | https://openxlab.org.cn/home
25 | ```
26 | 
27 | b. Install the dependent libraries as follows:
28 | 
29 | * Install the openxlab dependent libraries.
30 |   ```shell
31 |     pip install openxlab
32 |   ```
33 | * Obtain the Access Key and Secret Key on the OpenXLab website by clicking the button of Account Security
34 | * Login the OpenXLab using the Access Key and Secret Key
35 |   ```shell
36 |     openxlab login
37 |   ```
38 | 
39 | c. Download the SimChart9K dataset by performing the following command:
40 | ```shell
41 | openxlab dataset get --dataset-repo  Lonepic/SimChart9K
42 | ```
43 | 
44 | ## t-SNE comparisons with Real Chart Datasets
45 | 
46 | <p align="center">
47 |   <img src="images/t-SNE_a.png" width="62%">
48 |   <div>Feature Distribution using t-SNE of Real Datasets.</div>
49 | </p>
50 | 
51 | 
52 | <p align="center">
53 |   <img src="images/t-SNE_b.png" width="62%">
54 |   <div>Feature Distribution using t-SNE of both Real Datasets and SimChart9K.</div>
55 | </p>
56 | 
57 | ## Visualization Exapmles
58 | 
59 | <p align="center">
60 |   <img src="images/multi_task_1.png" width="85%">
61 |   <div>Visualization results using the proposed StructChart on different chart-related reasoning tasks including Question Answering (QA), Summarization, and Redrawing.</div>
62 | </p>
63 | 
64 | <p align="center">
65 |   <img src="images/multi_task_2.png" width="85%">
66 |   <div>Visualization results using the proposed StructChart on different chart-related reasoning tasks including Question Answering (QA), Summarization, and Redrawing.</div>
67 | </p>
68 | 
69 | ## Citation
70 | Please consider citing our work if this dataset is helpful for your research:
71 | 
72 | ```
73 | @article{xia2023structchart,
74 |   title={StructChart: Perception, Structuring, Reasoning for Visual Chart Understanding},
75 |   author={Xia, Renqiu and Zhang, Bo and Peng, Haoyang and Ye, Hancheng and Yan, Xiangchao and Ye, Peng and Shi, Botian and Yan, Junchi and Qiao, Yu},
76 |   journal={arXiv preprint arXiv:2309.11268},
77 |   year={2023}
78 | }
79 | ```
80 | 


--------------------------------------------------------------------------------
/images/SC.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Alpha-Innovator/SimChart9K/4afc2c5385d8bfd11f64e1356c25bffd20369114/images/SC.png


--------------------------------------------------------------------------------
/images/multi_task_1.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Alpha-Innovator/SimChart9K/4afc2c5385d8bfd11f64e1356c25bffd20369114/images/multi_task_1.png


--------------------------------------------------------------------------------
/images/multi_task_2.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Alpha-Innovator/SimChart9K/4afc2c5385d8bfd11f64e1356c25bffd20369114/images/multi_task_2.png


--------------------------------------------------------------------------------
/images/t-SNE_a.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Alpha-Innovator/SimChart9K/4afc2c5385d8bfd11f64e1356c25bffd20369114/images/t-SNE_a.png


--------------------------------------------------------------------------------
/images/t-SNE_b.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Alpha-Innovator/SimChart9K/4afc2c5385d8bfd11f64e1356c25bffd20369114/images/t-SNE_b.png


--------------------------------------------------------------------------------