Access Face Vision

Face detection and recognition Application.

With pip

• Installation

pip install access-face-vision

• Training/Creating FaceGroup

python -m access_face_vision --mode train --img_dir samples/celebrities --face_group celebrities

# Directory structure
# **/Images/
#          A/
#           A_K_01.jpg
#           A_K_02.jpg
#          B/
#           B_S_01.jpg
#           B_S_02.jpg

• Running Inferences

# Live video feed
python -m access_face_vision --mode live-video --camera-index 0 --camera_wait 30 --face_group celebrities

# server mode
python -m access_face_vision --mode server --face_group_dir ./

Use access-client to make requests to this server

Docker image

• Docker Image build

docker build -t access_face_vision:latest .

• Docker run

# we will use it as root directory for access_face_vision application
mkdir -p accessai/afv

# Start server
docker run -v $(pwd)/afv:/accessai/afv python -m access_face_vision --mode server

# Start camera feed processor
docker run -v $(pwd)/afv:/accessai/af v python -m access_face_vision --mode server

Contribution

Contributions are welcome. Feel free to raise PRs! with any improvements.

Credit

Face Encoder model: https://github.com/nyoki-mtl/keras-facenet