/face_detection

Primary LanguagePython

Face Mesh Detection

Requirements

Setup env:

virtualenv venv
source venv/bin/activate

Library:

  • MediaPipe
pip install mediapipe
pip install mediapipe-silicon

mediapipe>=0.8.9; platform_system != "Darwin" and platform.machine != 'arm64'
mediapipe-silicon>=0.8.9; platform_system == "Darwin" and platform.machine == 'arm64'
  • OpenCv
pip install opencv-python
  • numpy
pip install numpy
  • os
  • json
  • argparse Facial landmark recognition for 68 face landmark points:
  • we download from https://github.com/italojs/facial-landmarks-recognition/blob/master/shape_predictor_68_face_landmarks.dat
  • we save path ./Face Mesh Detection Example/shape_predictor_68_face_landmarks.dat

Face Mesh Detection Basic

Face detection and Face mesh detection:

  • Running following command to recognite face in image:
py face_detection.py
  • And running following command to get face mesh in image:
py face_mesh_detection.py
  • Running following command to help:
py face_detection.py -h

py face_mesh_detection.py -h
  • The image is in Photos folder and face detection results will be saved in FaceDetectionSavedImage and face mesh results will be saved in FaceMeshDetectionSavedImage. We can change image and save dir:
usage: face_detection.py [-h] [--sav SAV] [-dir DIR_IMAGE]

face detection

options:
  -h, --help            show this help message and exit
  --sav SAV             save dir
  -dir DIR_IMAGE, --dir-image DIR_IMAGE
                        folder of image

usage: face_mesh_detection.py [-h] [--sav SAV] [-dir DIR_IMAGE]

face detection

options:
  -h, --help            show this help message and exit
  --sav SAV             save dir
  -dir DIR_IMAGE, --dir-image DIR_IMAGE
                        folder of image

Face Detection Example

Delaunay Triangulation:

  • Running following command with save_dir (default: Delaunay triangle) and source_image (default: man.jpg):
py delaunay_triangulation.py [-h] [--sav SAV] [-src SOURCE_IMAGE]

Face swap 468 face landmark points:

  • Running following command with save_dir (default: face_swapped) and source_image (default: long.jpg), destination_image (default: man_2.jpg):
py face_swap_468_landmark_points.py [-h] [--sav SAV] [-src SOURCE_IMAGE] [-dst DESTINATION_IMAGE]

Face swap 68 face landmark points:

  • Running following command with save_dir (default: face_swapped) and source_image (default: long.jpg), destination_image (default: man_2.jpg):
py face_swapping.py [-h] [--sav SAV] [-src SOURCE_IMAGE] [-dst DESTINATION_IMAGE]

Face mesh detection and stick flag:

  • Running following command with save_dir (default: Stick Flag Save Image) and source_image (default: DSC_6730.JPG), flag (default: vietnam_flag,jpg):
py face_mesh_detection.py [-h] [-src SOURCE_IMAGE] [-flag FLAG] [--sav SAV]

Stick flag to face

options:
  -h, --help            show this help message and exit
  -src SOURCE_IMAGE, --source-image SOURCE_IMAGE
                        face is sticked
  -flag FLAG            image to stick into face
  --sav SAV             save dir

Geometric Art:

  • Running following command with save_dir (default: Geometric Art Save Image), source_image (default: long.jpg) and model (default: 1):
py geomrtric_art.py [-h] [-src SOURCE_IMAGE] [--model MODEL] [--sav SAV]

Geometric art

options:
  -h, --help            show this help message and exit
  -src SOURCE_IMAGE, --source-image SOURCE_IMAGE
                        face is make up
  --model MODEL         algorithms to get color: 1 is min and 0 is max
  --sav SAV             save dir