The task in this project is to match a face image without any occlusion with a face image covered with one of these masking approaches. This repo is a solution of summer challenge given by IITJ.
- Clone this repository
git clone https://github.com/Garvit-32/Face-recognition-in-presence-of-Mask.git
- Install all the dependencies
pip install -r requirements.txt
- Download and paste the pretrained weight of face detection model in ./face_detection/weights/ using the link
- Download the hand made dataset using the link ~ 110 MB
- Extract and paste it in project root directory
- (optional) You can use your dataset also but it follow a specific structure of directory
-dataset/
-person_name_1/
-image_01
-image_02
-person_name_2/
-image_01
-image_02
- Run generate_database python file to virtually append mask on person face in database
python generate_database.py
- Train the model
python train.py
- Extract face in image using extract_face.py edit test_data_path = "to_your_test_path"
python extract_face.py
- Test your image
python test.py