This project includes face detection / alignment / 5 or 68 or 106 landmarks detection / face attributes detection including age+gender+race+smile / face orientation detection / face recognition.
Basically it covers all fields of face tech except for face living detection which is also really important in the whole chain. But as it usually runs on mobile devices, so it's not contained in this repo which focused on backend usage.
tensorflow
caffe for python binding
numpy
opencv-python
dlib
scikit-learn
scikit-video
nudged
I wrote a face server backend based on flask. Just run ./run.sh to start it and then you can visit the api provided.
More details can be found in ./backend_api/.
ATTENSION for below settings:
CUDA_VISIBLE_DEVICES/PYTHONPATH/port setting, you may need modify this according to your env.
Any problem is welcome to report to issue or contact buptmsg@gmail.com.
| test image | face detection | landmark detection | age | gender | race | smile | celebrity |
|---|---|---|---|---|---|---|---|
 |
 |
 |
24.47 | female | asian | smile | Fan Bingbing |
 |
 |
 |
60.15 | male | asian | calm | Guo Degang |
 |
 |
 |
26.06 | female | asian | smile | XinYuanJieYi |
 |
 |
 |
54.52 | male | african | calm | AoBaMa |
Wo De Nan Shen AAAAAAAAAAA!
Wo De Nv Shen EEEEEEEEEEE!