Skip to content

Latest commit

 

History

History
46 lines (34 loc) · 1.29 KB

README.md

File metadata and controls

46 lines (34 loc) · 1.29 KB
Raw

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>.