Face Detector Pool
- Implement common face detection algorithms.
Requirements
- Python 3.6
- OpenCV 3
- Dlib
- Tensorflow 1.4
Current methods
- Dlib
- MTCNN
Usage
- Use scripts
python run_face_analyze.py [--detector [method]]
[--input [image file or folder]]
[--viz [visualize or not]]
For example, python run_face_analyze.py --detector dlib --input Yuniko --viz True
-
Use high level class
-
Create FaceAnalyzer() instance with a detector.
-
Run full_analyze() will return detected faces and related facial landmarks.
-
-
Use detector api
- detect_faces() return face bounding boxes.
- detect_facial_landmarks() return facial landmarks.
Outputs
- Detected face will be cropped and saved into
output
folder. - Related landmarks will be saved into
face_metadata.json
with<key, value> = <img_path, landmarks>
.