face dataset
>> python ict_download.py
download it to project directory shape_predictor_68_face_landmarks.dat file
make image meta file.
run
python pyqt.py
and then open meta file directory.
now you can edit landmark detection.
60 images
15 neutral pose num : 1 - 0 degree pose num : 6 - yaw -90 ~ 90 (per 30 degree) : this mean left and right head num : 4 - pitch-30 ~ 30 (per 15 degree) num : 4 - roll -30 ~ 30 (per 15 degree)
non-rigid pose(expression) 3 yaw angle -30, 0, 30 mouth stretch
- smile
- brow raise
- disgust
- squeeze left eye
- squeeze right eye
- anger
- jaw left
- jaw right
- grin
- chin raise
- lip pucker
- lip funnel
- cheek blowing
- eyes closed
[basis_w, expr_1_w expr_2_w .... expr_n_w ]