Face_recognition

This project is based on face recognition on live camera. In this project I have used facenet to face recognition. Haar cascade for face detection. Integrate both concept to recognition face on own dataset.

Install prerequisites

pip install -r requirements.txt

Note:- One can use any version of tensorflow with cuda compatible.

How to train on own dataset

• Download pretrained weight folder past into repo. folder.
• Create datasets folder in same repo. folder.
• put images(Your sown dataset) in classes-wise(name-wise) folder.

Run following command which can create classifier.pkl in model folder:

python src/classifier.py TRAIN ./datasets ./model/20180402-114759.pb ./model/classifier.pkl --batch_size 1

Run trained model on webcam:

python src/video.py CLASSIFY ./datasets ./model/20180402-114759.pb ./model/classifier.pkl --batch_size 1