├── .gitignore ├── packages.txt ├── db ├── Jack Ma.jpg ├── Bill Gates.png ├── Elon Musk.jpg ├── Jeff Bezos.jpg ├── Mark Zuckerberg.jpg └── Warren Buffett.jpg ├── encoded_faces.pickle ├── test pictures ├── Elon Musk.jpg ├── Jack Ma.jpg ├── Bill Gates.png ├── Jeff Bezos.jpg ├── Warren Buffett.jpg ├── mark_jack ma.jpg ├── Mark Zuckerberg.jpg ├── bill gates_elon musk_ jeff bezos_1.jpg └── Elon Musk_Bill Gates_Jeff Bezos_Mark Zuckerberg.jpg ├── environment.yml ├── Attendance.csv ├── Training.py ├── Preparing_test_online.py ├── streamlit_local_app_bussines_ready.py ├── Preparing_local.py ├── streamlit_test_app_online.py └── README.md /.gitignore: -------------------------------------------------------------------------------- 1 | __pycache__/ -------------------------------------------------------------------------------- /packages.txt: -------------------------------------------------------------------------------- 1 | cmake 2 | libgtk-3-dev 3 | freeglut3-dev 4 | -------------------------------------------------------------------------------- /db/Jack Ma.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/AbdassalamAhmad/Attendance_System/HEAD/db/Jack Ma.jpg -------------------------------------------------------------------------------- /db/Bill Gates.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/AbdassalamAhmad/Attendance_System/HEAD/db/Bill Gates.png -------------------------------------------------------------------------------- /db/Elon Musk.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/AbdassalamAhmad/Attendance_System/HEAD/db/Elon Musk.jpg -------------------------------------------------------------------------------- /db/Jeff Bezos.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/AbdassalamAhmad/Attendance_System/HEAD/db/Jeff Bezos.jpg -------------------------------------------------------------------------------- /db/Mark Zuckerberg.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/AbdassalamAhmad/Attendance_System/HEAD/db/Mark Zuckerberg.jpg -------------------------------------------------------------------------------- /db/Warren Buffett.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/AbdassalamAhmad/Attendance_System/HEAD/db/Warren Buffett.jpg -------------------------------------------------------------------------------- /encoded_faces.pickle: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/AbdassalamAhmad/Attendance_System/HEAD/encoded_faces.pickle -------------------------------------------------------------------------------- /test pictures/Elon Musk.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/AbdassalamAhmad/Attendance_System/HEAD/test pictures/Elon Musk.jpg -------------------------------------------------------------------------------- /test pictures/Jack Ma.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/AbdassalamAhmad/Attendance_System/HEAD/test pictures/Jack Ma.jpg -------------------------------------------------------------------------------- /test pictures/Bill Gates.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/AbdassalamAhmad/Attendance_System/HEAD/test pictures/Bill Gates.png -------------------------------------------------------------------------------- /test pictures/Jeff Bezos.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/AbdassalamAhmad/Attendance_System/HEAD/test pictures/Jeff Bezos.jpg -------------------------------------------------------------------------------- /test pictures/Warren Buffett.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/AbdassalamAhmad/Attendance_System/HEAD/test pictures/Warren Buffett.jpg -------------------------------------------------------------------------------- /test pictures/mark_jack ma.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/AbdassalamAhmad/Attendance_System/HEAD/test pictures/mark_jack ma.jpg -------------------------------------------------------------------------------- /test pictures/Mark Zuckerberg.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/AbdassalamAhmad/Attendance_System/HEAD/test pictures/Mark Zuckerberg.jpg -------------------------------------------------------------------------------- /test pictures/bill gates_elon musk_ jeff bezos_1.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/AbdassalamAhmad/Attendance_System/HEAD/test pictures/bill gates_elon musk_ jeff bezos_1.jpg -------------------------------------------------------------------------------- /test pictures/Elon Musk_Bill Gates_Jeff Bezos_Mark Zuckerberg.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/AbdassalamAhmad/Attendance_System/HEAD/test pictures/Elon Musk_Bill Gates_Jeff Bezos_Mark Zuckerberg.jpg -------------------------------------------------------------------------------- /environment.yml: -------------------------------------------------------------------------------- 1 | name: attendance 2 | channels: 3 | - conda-forge 4 | dependencies: 5 | - dlib 6 | - pandas 7 | - numpy 8 | - face_recognition 9 | - pip: 10 | - opencv-python 11 | -------------------------------------------------------------------------------- /Attendance.csv: -------------------------------------------------------------------------------- 1 | Name,Arrive_Time,Date,Late_penalty 2 | Bill Gates,07:41:57,2022/03/15,0 3 | Jack Ma,07:42:01,2022/03/15,0 4 | Warren Buffett,07:42:04,2022/03/15,0 5 | Jeff Bezos,07:42:07,2022/03/15,0 6 | Elon Musk,07:42:10,2022/03/15,0 -------------------------------------------------------------------------------- /Training.py: -------------------------------------------------------------------------------- 1 | import cv2 2 | import face_recognition 3 | import os 4 | # import streamlit as st 5 | # import pickle 6 | # Declaring variables 7 | path = "db" 8 | 9 | def training(path): 10 | images = os.listdir(path) 11 | encoded_trains =[] 12 | 13 | for image in images: 14 | train_img = face_recognition.load_image_file(f"db/{image}") 15 | train_img = cv2.cvtColor(train_img,cv2.COLOR_BGR2RGB) 16 | encoded_trains.append(face_recognition.face_encodings(train_img)[0]) 17 | return encoded_trains, images 18 | 19 | 20 | # encoded_trains, images = training(path) 21 | # st.write(images) 22 | # output_file = 'encoded_faces.pickle' 23 | # with open(output_file, 'wb') as f_out: 24 | # pickle.dump(encoded_trains, f_out) 25 | -------------------------------------------------------------------------------- /Preparing_test_online.py: -------------------------------------------------------------------------------- 1 | import cv2 2 | import face_recognition 3 | import os 4 | import time 5 | import pandas as pd 6 | import streamlit as st 7 | # Declaring variables 8 | path = "db" 9 | scale = 2 10 | 11 | 12 | def late_penalty(arrive_time): 13 | penalty = 0 14 | if int(arrive_time[0:2])> 8: 15 | penalty = 10 16 | return penalty 17 | 18 | 19 | def markattendance(person_name, new_date): 20 | with open("Attendance.csv",'r+') as f: 21 | lines = f.readlines() 22 | name_list=[] 23 | now = time.localtime() 24 | date = new_date 25 | 26 | for line in lines: 27 | entry = line.split(',') 28 | if len(entry)>1: 29 | if entry[2] == date: 30 | name_list.append(entry[0]) 31 | 32 | if person_name not in name_list: 33 | arrive_time = time.strftime("%H:%M:%S", now) 34 | penalty = late_penalty(arrive_time) 35 | f.writelines(f'\n{person_name},{arrive_time},{date},{penalty}') 36 | 37 | f = open("Attendance.csv",'r',encoding = 'utf-8') 38 | df = pd.read_csv(f) 39 | return df 40 | 41 | 42 | def prepare_test_img(test_img): 43 | test_img = face_recognition.load_image_file(test_img) 44 | #test_img = cv2.cvtColor(test_img,cv2.COLOR_BGR2RGB) 45 | test_img_small = cv2.resize(test_img,(0,0),None,0.5,0.5) 46 | 47 | face_test_locations = face_recognition.face_locations(test_img_small, model = "hog") 48 | encoded_tests = face_recognition.face_encodings(test_img_small) 49 | return test_img, encoded_tests, face_test_locations 50 | 51 | 52 | def test(encoded_tests, face_test_locations, test_img, encoded_trains, new_date): 53 | images = os.listdir(path) 54 | name_indices = [] 55 | df ="No Faces Found" 56 | for encoded_test, face_test_location in zip(encoded_tests, face_test_locations): 57 | results = face_recognition.compare_faces(encoded_trains,encoded_test,tolerance=0.6) 58 | # face_distances = face_recognition.face_distance(encoded_trains,encoded_test) 59 | # st.write(face_distances) 60 | # st.write(images) 61 | if True in results: 62 | name_index = results.index(True) 63 | name_indices.append(name_index) 64 | for count, image in enumerate(images): 65 | if count == name_index: 66 | person_name = image.split(".")[0] 67 | top_left, bottom_right = (face_test_location[3]*scale, face_test_location[0]*scale) ,(face_test_location[1]*scale, face_test_location[2]*scale) 68 | cv2.rectangle(test_img,(top_left),(bottom_right),(255,0,255),2) 69 | cv2.rectangle(test_img,(bottom_right),(top_left[0], bottom_right[1]+30),(255,0,255),cv2.FILLED) 70 | cv2.putText(test_img,person_name,(top_left[0]+6,bottom_right[1]+25),cv2.FONT_HERSHEY_COMPLEX,1,(255,255,255),1) 71 | df=markattendance(person_name, new_date) 72 | else: 73 | top_left, bottom_right = (face_test_location[3]*scale, face_test_location[0]*scale) ,(face_test_location[1]*scale, face_test_location[2]*scale) 74 | cv2.rectangle(test_img,(top_left),(bottom_right),(255,0,255),2) 75 | cv2.rectangle(test_img,(bottom_right),(top_left[0], bottom_right[1]+30),(255,0,255),cv2.FILLED) 76 | cv2.putText(test_img,"UNKNOWN",(top_left[0]+6,bottom_right[1]+25),cv2.FONT_HERSHEY_COMPLEX,1,(255,255,255),1) 77 | f = open("Attendance.csv",'r',encoding = 'utf-8') 78 | df = pd.read_csv(f) 79 | return (df) -------------------------------------------------------------------------------- /streamlit_local_app_bussines_ready.py: -------------------------------------------------------------------------------- 1 | import streamlit as st 2 | import time 3 | import cv2 4 | import pickle 5 | import face_recognition 6 | from Preparing_local import prepare_test_img, test 7 | 8 | t0= time.time() 9 | # print("Hello") 10 | # Declaring variables 11 | path = "db" 12 | 13 | 14 | def main(): 15 | # Loading the mode 16 | #@st.cache 17 | def load_model(): 18 | with open ('encoded_faces.pickle', 'rb') as f_in: 19 | encoded_trains = pickle.load(f_in) 20 | return encoded_trains 21 | encoded_trains = load_model() 22 | 23 | # Start of the project 24 | st.title("Attendance_Project") 25 | st.sidebar.title("What to do") 26 | app_mode = st.sidebar.selectbox("Choose the app mode", 27 | ["Attend from image", "Attend using camera", "Training", "Attend Live"]) 28 | 29 | 30 | if app_mode == "Attend from image": 31 | attendance_file = st.file_uploader("Choose attendance file",type =['csv']) 32 | uploaded_file = st.file_uploader("Upload a picture of a person to make him attend", type=['jpg', 'jpeg', 'png']) 33 | if attendance_file is not None and uploaded_file is not None: 34 | 35 | test_img, encoded_tests, face_test_locations = prepare_test_img(uploaded_file) 36 | df = test(encoded_tests, face_test_locations, test_img, encoded_trains, attendance_file) 37 | t1 = time.time() - t0 38 | 39 | st.write("Time elapsed: ", t1) 40 | # test_img = cv2.resize(test_img,(0,0),None,0.50,0.50) 41 | st.image(test_img) 42 | st.write(df) 43 | 44 | 45 | elif app_mode == "Attend using camera": 46 | attendance_file = st.file_uploader("Choose attendance file",type =['csv']) 47 | picture = st.camera_input("Take a picture") 48 | if picture is not None and attendance_file is not None: 49 | 50 | test_img, encoded_tests, face_test_locations = prepare_test_img(picture) 51 | df = test(encoded_tests, face_test_locations, test_img, encoded_trains, attendance_file) 52 | t1 = time.time() - t0 53 | 54 | st.write("Time elapsed: ", t1) 55 | #test_img = cv2.resize(test_img,(0,0),None,0.50,0.50) 56 | st.image(test_img) 57 | st.write(df) 58 | 59 | 60 | elif app_mode == "Training": 61 | st.subheader('Training Steps:') 62 | st.markdown("1. Get a photo of every employee with **only one face** in the picture.") 63 | st.markdown('2. Put all the photos in the **db** folder') 64 | st.markdown("3. Press **Train The Model** Button") 65 | 66 | if st.button("Train The Model"): 67 | import Training 68 | encoded_trains, images = Training.training(path) 69 | st.write(images) 70 | st.write(len(encoded_trains)) 71 | output_file = 'encoded_faces.pickle' 72 | 73 | with open(output_file, 'wb') as f_out: 74 | pickle.dump(encoded_trains, f_out) 75 | 76 | 77 | elif app_mode == "Attend Live": 78 | st.title("Webcam Live Feed") 79 | attendance_file = st.file_uploader("Choose attendance file",type =['csv']) 80 | 81 | if attendance_file is not None: 82 | run = st.checkbox('Run') 83 | FRAME_WINDOW = st.image([]) 84 | camera = cv2.VideoCapture(0) 85 | 86 | while run: 87 | _, test_img = camera.read() 88 | test_img = cv2.cvtColor(test_img, cv2.COLOR_BGR2RGB) 89 | test_img_small = cv2.resize(test_img,(0,0),None,0.5,0.5) 90 | 91 | face_test_locations = face_recognition.face_locations(test_img_small, model = "hog") 92 | encoded_tests = face_recognition.face_encodings(test_img_small) 93 | df = test(encoded_tests, face_test_locations, test_img, encoded_trains, attendance_file) 94 | #st.image(test_img) 95 | FRAME_WINDOW.image(test_img) 96 | #st.write(df) 97 | else: 98 | st.write('Stopped') 99 | 100 | 101 | if __name__=='__main__': 102 | main() -------------------------------------------------------------------------------- /Preparing_local.py: -------------------------------------------------------------------------------- 1 | import cv2 2 | import face_recognition 3 | import os 4 | import time 5 | import pandas as pd 6 | import streamlit as st 7 | 8 | # Declaring variables 9 | path = "db" 10 | scale = 2 11 | 12 | def late_penalty(arrive_time): 13 | penalty = 0 14 | if int(arrive_time[0:2])> 8: 15 | penalty = 10 16 | return penalty 17 | 18 | def markattendance(person_name, attendance_file): 19 | 20 | with open(attendance_file.name,'r+') as f: 21 | lines = f.readlines() 22 | name_list=[] 23 | now = time.localtime() 24 | date = time.strftime("%Y/%m/%d", now) 25 | 26 | for line in lines: 27 | entry = line.split(',') 28 | if len(entry)>1: 29 | if entry[2] == date: 30 | name_list.append(entry[0]) 31 | 32 | if person_name not in name_list: 33 | arrive_time = time.strftime("%H:%M:%S", now) 34 | penalty = late_penalty(arrive_time) 35 | f.writelines(f'\n{person_name},{arrive_time},{date},{penalty}') 36 | 37 | f = open(attendance_file.name,'r',encoding = 'utf-8') 38 | df = pd.read_csv(f) 39 | return df 40 | 41 | 42 | def prepare_test_img(test_img): 43 | test_img = face_recognition.load_image_file(test_img) 44 | #test_img = cv2.cvtColor(test_img,cv2.COLOR_BGR2RGB) 45 | test_img_small = cv2.resize(test_img,(0,0),None,0.5,0.5) 46 | 47 | face_test_locations = face_recognition.face_locations(test_img_small, model = "hog") 48 | encoded_tests = face_recognition.face_encodings(test_img_small) 49 | return test_img, encoded_tests, face_test_locations 50 | 51 | 52 | def test(encoded_tests, face_test_locations, test_img, encoded_trains, attendance_file): 53 | images = os.listdir(path) 54 | name_indices = [] 55 | df ="No Faces Found" #for handling an error when no faces detected 56 | 57 | for encoded_test, face_test_location in zip(encoded_tests, face_test_locations): 58 | results = face_recognition.compare_faces(encoded_trains,encoded_test,tolerance=0.49) 59 | # face_distances = face_recognition.face_distance(encoded_trains,encoded_test) 60 | # st.write(min(face_distances)) 61 | 62 | if True in results: 63 | name_index = results.index(True) 64 | name_indices.append(name_index) 65 | 66 | for count, image in enumerate(images): 67 | if count == name_index: 68 | person_name = image.split(".")[0] 69 | top_left, bottom_right = (face_test_location[3]*scale, face_test_location[0]*scale) ,(face_test_location[1]*scale, face_test_location[2]*scale) 70 | cv2.rectangle(test_img,(top_left),(bottom_right),(255,0,255),2) 71 | cv2.rectangle(test_img,(bottom_right),(top_left[0], bottom_right[1]+30),(255,0,255),cv2.FILLED) 72 | cv2.putText(test_img,person_name,(top_left[0]+6,bottom_right[1]+25),cv2.FONT_HERSHEY_COMPLEX,1,(255,255,255),1) 73 | df=markattendance(person_name, attendance_file) 74 | 75 | else: 76 | top_left, bottom_right = (face_test_location[3]*scale, face_test_location[0]*scale) ,(face_test_location[1]*scale, face_test_location[2]*scale) 77 | cv2.rectangle(test_img,(top_left),(bottom_right),(255,0,255),2) 78 | cv2.rectangle(test_img,(bottom_right),(top_left[0], bottom_right[1]+30),(255,0,255),cv2.FILLED) 79 | cv2.putText(test_img,"UNKNOWN",(top_left[0]+6,bottom_right[1]+25),cv2.FONT_HERSHEY_COMPLEX,1,(255,255,255),1) 80 | f = open(attendance_file.name,'r',encoding = 'utf-8') 81 | df = pd.read_csv(f) 82 | 83 | # this code was to put attendances in a dictionary but now i'm using pandas 84 | # attendance_list = [False for i in range(len(results))] 85 | # for i in name_indices: 86 | # attendance_list[i] = True 87 | 88 | # for image in images: 89 | # names.append(image.split(".")[0]) 90 | 91 | # ans = {} 92 | # for i, name in enumerate (names): 93 | # ans[name] = attendance_list[i] 94 | return (df) 95 | 96 | 97 | 98 | 99 | 100 | 101 | 102 | -------------------------------------------------------------------------------- /streamlit_test_app_online.py: -------------------------------------------------------------------------------- 1 | import streamlit as st 2 | import time 3 | import pickle 4 | import os 5 | from Preparing_test_online import prepare_test_img, test 6 | 7 | t0= time.time() 8 | #print("Hello") 9 | 10 | # Declaring variables 11 | path = "db" 12 | test_path = "test pictures" 13 | 14 | def main(): 15 | # Loading the model 16 | #@st.cache 17 | def load_model(): 18 | with open ('encoded_faces.pickle', 'rb') as f_in: 19 | encoded_trains = pickle.load(f_in) 20 | return encoded_trains 21 | encoded_trains = load_model() 22 | 23 | # Start of the project 24 | st.title("Attendance_Project") 25 | st.sidebar.title("What to do") 26 | app_mode = st.sidebar.selectbox("Choose the app mode", 27 | ["Demo app", "Attend from uploading image", "Attend using camera (photo mode)", "Training"]) 28 | 29 | if app_mode == "Demo app": 30 | st.sidebar.write(" ------ ") 31 | photos = [] 32 | images = os.listdir(test_path) 33 | 34 | for image in images: 35 | filepath = os.path.join(test_path, image) 36 | photos.append(image) 37 | 38 | option = st.sidebar.selectbox('Please select a sample image, then click Attend button', photos) 39 | pressed = st.sidebar.button('Attend') 40 | 41 | if pressed: 42 | st.sidebar.write('Please wait for the magic to happen!') 43 | pic = os.path.join(test_path, option) 44 | test_img, encoded_tests, face_test_locations = prepare_test_img(pic) 45 | 46 | ############----for trying only----------########## 47 | now = time.localtime() 48 | date = time.strftime("%Y/%m/%d", now) 49 | new_date = st.text_input('For Trying purposes you can put any date to test the program', date) 50 | ############----end trying----------########## 51 | 52 | df = test(encoded_tests, face_test_locations, test_img, encoded_trains, new_date) 53 | st.image(test_img) 54 | st.write(df) 55 | 56 | elif app_mode == "Attend from uploading image": 57 | uploaded_file = st.file_uploader("Upload a picture of a person to make him attend", type=['jpg', 'jpeg', 'png']) 58 | if uploaded_file is not None: 59 | 60 | test_img, encoded_tests, face_test_locations = prepare_test_img(uploaded_file) 61 | ############----for trying only----------########## 62 | now = time.localtime() 63 | date = time.strftime("%Y/%m/%d", now) 64 | new_date = st.text_input('For Trying purposes you can put any date to test the program', date) 65 | ############----end trying----------########## 66 | 67 | df = test(encoded_tests, face_test_locations, test_img, encoded_trains, new_date) 68 | t1 = time.time() - t0 69 | 70 | st.write("Time elapsed: ", t1) 71 | st.image(test_img) 72 | st.write(df) 73 | 74 | 75 | elif app_mode == "Attend using camera (photo mode)": 76 | picture = st.camera_input("Take a picture of yourself to attend") 77 | if picture is not None: 78 | 79 | test_img, encoded_tests, face_test_locations = prepare_test_img(picture) 80 | ############----for trying only----------########## 81 | now = time.localtime() 82 | date = time.strftime("%Y/%m/%d", now) 83 | new_date = st.text_input('For Trying purposes you can put any date to test the program', date) 84 | ############----end trying----------########## 85 | df = test(encoded_tests, face_test_locations, test_img, encoded_trains, new_date) 86 | t1 = time.time() - t0 87 | 88 | st.write("Time elapsed: ", t1) 89 | st.image(test_img) 90 | st.write(df) 91 | 92 | 93 | elif app_mode == "Training": 94 | st.subheader('Training Steps:') 95 | st.markdown("1. Get a photo of every employee with **only one face** in the picture.") 96 | st.markdown('2. Put all the photos in the **db** folder') 97 | st.markdown("3. Press **Train The Model** Button") 98 | if st.button("Train The Model"): 99 | import Training 100 | encoded_trains, images = Training.training(path) 101 | st.write(images) 102 | st.write(len(encoded_trains)) 103 | output_file = 'encoded_faces.pickle' 104 | with open(output_file, 'wb') as f_out: 105 | pickle.dump(encoded_trains, f_out) 106 | 107 | 108 | if __name__=='__main__': 109 | main() -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | [![Open in Streamlit](https://static.streamlit.io/badges/streamlit_badge_black_white.svg)](https://share.streamlit.io/abdassalamahmad/attendance_system/main/streamlit_test_app_online.py) 2 | # Attendance-System Using Face-Recognition in Real-Time 3 | 4 | ## Model Demo 5 | 6 | https://user-images.githubusercontent.com/83673888/158333701-9f2cffd8-cd91-496b-98f7-94702bf4bcb6.mp4 7 | 8 | ## How To Try This Model 9 | There are two versions (Online and Offline) 10 | 11 | | | Online | Offline | 12 | | --------------------------------| :--------------------------------------------: | :------------------------------------------------------------: | 13 | | Training | - Only the host (me) can train new faces. | - When you clone the repo you can train new faces as you want. | 14 | | Attend from Uploading Photo |
  • - [x]
    • |
  • - [x]
    • | 15 | | Attend from Camera (Photo Mode) |
  • - [x]
    • |
  • - [x]
    • | 16 | | Attend Live |
  • - [ ]
    • |
  • - [x]
    • | 17 | 18 | #### **Online Version:**
    you can try it from [here](https://share.streamlit.io/abdassalamahmad/attendance_system/main/streamlit_test_app_online.py) 19 | #### **Offline Version:**
    Follow these steps to try it: 20 | 1. Clone this repo to get all the code and pre-trained model(pickle_file). 21 | 2. Change current directory into the cloned repo folder. 22 | 3. Install all of the libraries from [environment.yml](https://github.com/AbdassalamAhmad/Attendance_System/blob/main/environment.yml) file by using these commands. 23 | ``` 24 | conda env create -f environment.yml 25 | conda activate attendance 26 | ``` 27 | (optional step) to check if all libraries installed 28 | ``` 29 | conda env list 30 | ``` 31 | 4. Install all of the dependencies from [packages.txt](https://github.com/AbdassalamAhmad/Attendance_System/blob/main/packages.txt) using this command. 32 | - **Linux users**: cmake is a must. 33 | ``` 34 | sudo apt-get install cmake libgtk-3-dev freeglut3-dev 35 | ``` 36 | - **Windows users**:
    You need to install visual studio community version from [here](https://visualstudio.microsoft.com/downloads/) and make sure that cmake is checked when installing because it is a must. 37 | 5. Run this script [streamlit_local_app_bussines_ready.py](https://github.com/AbdassalamAhmad/Attendance_System/blob/main/streamlit_local_app_bussines_ready.py) to try the offline version by running this code in the cloned repo directory after installing all dependencies. 38 | ```py 39 | streamlit run streamlit_local_app_bussines_ready.py 40 | ``` 41 | 42 | Training : 43 | ---------- 44 | To train the model on different faces, do the following: 45 | 1. Get a photo that contains one person and rename it to the person's name and put it in the `db` folder like this picture.
    46 | ![image](https://user-images.githubusercontent.com/83673888/158378862-30e4cce9-a737-4079-ae8c-99a013ea7460.png)
    47 | 2. Repeat that for as many faces as you want. 48 | 3. Finally run the [streamlit_local_app_bussines_ready.py](https://github.com/AbdassalamAhmad/Attendance_System/blob/main/streamlit_local_app_bussines_ready.py) script, and put it on Training mode, and press the `Train The Model` button, then you can go for testing. 49 | 50 | Testing : 51 | --------- 52 | - You have three modes. The best one is **Live Attendance** (Real case scenario)
    53 | To run it do the following: 54 | 1. From the sidebar select `Attend Live` mode. 55 | 2. Select `Attendance.csv` file, which is a file to record the arrival_time, date, penalty of every attendant. 56 | 3. Check `run` box to start the program then show the camera faces of people you've trained (people in `db` folder) 57 | 58 | - You can also try attending from uploading a picture of your face and it will work as well. 59 | To run it do the following: 60 | 1. From the sidebar select `Attend from uploading image` mode. 61 | 2. Upload your image or drag & drop it and it will detect your face and make you attend **just like this picture**. 62 | ![image](https://user-images.githubusercontent.com/83673888/158464863-65775d07-0023-4e7e-b15a-f2a09052af35.png) 63 | Note: Of course this is for trying purposes the best model is `Attend Live` described above. 64 | 65 | ## Short Description of the Files 66 | 1. [Preparing_local.py](https://github.com/AbdassalamAhmad/Attendance_System/blob/main/Preparing_local.py): It has all of the functions necessary for the main offline version script to work. 67 | 2. [Preparing_test_online.py](https://github.com/AbdassalamAhmad/Attendance_System/blob/main/Preparing_test_online.py): very similar file to the previous file with only some changes to make the online version work properly. 68 | 3. [Training.py](https://github.com/AbdassalamAhmad/Attendance_System/blob/main/Training.py): a script for training the model to memorize faces stored in `db` folder. 69 | 4. [encoded_faces.pickle](https://github.com/AbdassalamAhmad/Attendance_System/blob/main/encoded_faces.pickle): The output file of `Training.py` script, It contains the encoded features of every face to compare it with new faces. 70 | 5. [environment.yml](https://github.com/AbdassalamAhmad/Attendance_System/blob/main/environment.yml) and [packages.txt](https://github.com/AbdassalamAhmad/Attendance_System/blob/main/packages.txt) : libraries and dependencies to make this project work. 71 | 6. [streamlit_local_app_bussines_ready.py](https://github.com/AbdassalamAhmad/Attendance_System/blob/main/streamlit_local_app_bussines_ready.py): Attend live or from uploading photo or take a photo then attend. 72 | 7. [streamlit_test_app_online.py](https://github.com/AbdassalamAhmad/Attendance_System/blob/main/streamlit_test_app_online.py): very similar file to the previous script but without the **Live Attendance Feature** and it works online right from this [link](https://share.streamlit.io/abdassalamahmad/attendance_system/main/streamlit_test_app_online.py). 73 | 74 | ## How This Model Work (what is going on under the hood) 75 | 1. Detecting all of the faces in the picture/video:- using HOG algorithm. This function do the work `face_recognition.face_locations`. 76 | 2. Transform the face to make the eyes and mouth centered after finding main face landmarks using face landmark estimation. 77 | 3. Encode the faces by generating 128 different measuremts per face (saving faces). This function do the training (encodings) `face_recognition.face_encodings` 78 | 4. Recognition:- comparing new faces from photo/video with the encoded faces in our dataset. This function `face_recognition.compare_faces` do the comparing and return a list of True and False. 79 | 5. Make the attendance :- `markattendance()` this function uses OpenCV library to annotate the faces and then add the name each detected face -based on the previous function return `face_recognition.compare_faces`- to the attendance list [Attendance.csv](https://github.com/AbdassalamAhmad/Attendance_System/blob/main/Attendance.csv). 80 | 81 | 82 | ## Resources and Note 83 | I have used these resources to build my project: 84 | 1. [medium article](https://medium.com/@ageitgey/machine-learning-is-fun-part-4-modern-face-recognition-with-deep-learning-c3cffc121d78): - It descripe in details how **face recognition** is really working under the hood. 85 | 2. [YouTube Tutorial](https://www.youtube.com/watch?v=sz25xxF_AVE): - **for attendance part** using OpenCV and HOG. It is basically based on the previous article. 86 | 3. I've written the code myself based on the video and enhanced some features like when attending live it was mainly designed to be used for **one day only.**
    87 | I changed the logic and make it work **forever**.
    88 | I've aslo added the penalty feature for people who are late to work, (they get a 10$ penalty if they came after 9:00 AM). 89 | 90 | If you like this project, I appreciate you starring this repo.
    91 | Please feel free to fork the content and contact me on my [LinkedIn account](https://www.linkedin.com/in/abdassalam-ahmad/) if you have any questions. 92 | 93 | 94 | 95 | 96 | 97 | 98 | ## To-Do List 99 | - [x] Penalty (10$) for comming late to work (After 9:00 AM). 100 | - [ ] Detect 3D faces only. (printed faces or rendered faces on screens shouldn't be detected). 101 | --------------------------------------------------------------------------------