grommy / face-detection-and-recognition

face recognition experiments using opencv and neural nets

Face recognition experiments using opencv and neural nets

Prerequisites:

• Gather train dataset (photos) Structure dataset in a following way:

data/photos/
├── Person_1
│   ├── Screen\ Shot\ 2019-03-04\ at\ 9.39.34\ AM.png
│   ├── Screen\ Shot\ 2019-03-04\ at\ 9.39.51\ AM.png
...
├── Person_2
│   ├── DSC_0462-3.jpg
│   ├── DSC_0463-2.jpg

• unix-like os
• python3 and pip installed
• Install libs
  • install opencv
  • install face-regognition lib
  • install other requirements pip install -r requirements.txt

To run:

• encode known faces to vectors:
  python encode_faces.py --dataset data/photos/ -e data/encodings/face_vec.pkl -d 'dnn'
• run main script python recognize_faces.py --detection 'dnn' --encodings data/encodings/face_vec.pkl --input 'live' --display 1