ONNX implementation of YOLOv5 and Siamese Network (ResNet100) with ArcFace loss for Face Detection and Recognition
face_recog_nhom.mp4
conda install faceid torchvision torchaudio pytorch-cuda=11.7 -c pytorch -c nvidia
conda activate faceid
pip install opencv-python
pip install onnxruntime==1.14.0
pip install onnxruntime-gpu==1.14.0
git clone https://github.com/PhucNDA/FaceID--YOLOV5.ArcFace
cd FaceID--YOLOV5.ArcFaceEnsuring the right data tree format
FaceID--YOLOV5.ArcFace
├── database_image
│ ├── profile1.png
| ├── profile2.png
| ├── profile3.png
| ├── ...
├── database_tensor
│ ├── profile1.npy
| ├── profile2.npy
| ├── profile3.npy
| ├── ...
database_image: containing image for each profile
database_tensor: containing vector feature extracted by pretrained backbone for each profile
Making a few key modifications to the YOLOv5 and optimize it for face detection. These modifications include adding a five-point landmark regression head, using a stem block at the input of the backbone, using smaller-size kernels in the SPP, and adding a P6 output in the PAN block. YOLOv5 is pretrained on WIDERFace datasets and we already converted it to the ONNX gpu runtime:
yolov5m-face.onnxPre-trained backbone ResNet100 weights on Glint360K which contains 17091657 images of 360232 individuals is available at : weights
Manually add new face images to folder:
database_imageFor fast precomputation, pre-extract database images to .npy tensor:
python feature_extraction.py --weight 'weights/backbone.pth' --path_database database_imagedatabase_tensorConvert backbone weight to ONNX implementation:
python converttoonnx.pyIt is implemented on ResNet100 backbone and SOTA ArcFace loss: paper
Following is the pipeline of ArcFace loss:
Transfer-learning by training new faces on custom datasets is not neccessary in face recognition algorithm. Using a pre-trained weight on large-scale dataset is enough for feature extraction backbone. Using asia-oriented dataset might lead to bias toward inference phase.
python detection_gpu.pyThis version is good enough for face recognition system. Adding threshold for Unknown classification depends on user-experience. The model performs well on tilted face and obscured face (facemask). More improvements will be carried out in the future.
face_recog_phuc.mp4
I want to express sincere thanks to my colleagues at University of Information Technology - UIT for their contribution to this project.
| Serial | Full name | Github | |
|---|---|---|---|
| 1 | Nguyễn Đức Anh Phúc | PhucNDA | phucnda@gmail.com |
| 2 | Huỳnh Viết Tuấn Kiệt | hiimking1509 | 20521494@gm.uit.edu.vn |
| 3 | Nguyễn Nhật Trường | truongnn | 20522087@gm.uit.edu.vn |
| 4 | Lại Chí Thiện | laichithien | 20520309@gm.uit.edu.vn |
| 5 | Lê Việt Thịnh | levietthinh | vietthinh101@gmail.com |
This project is carried out in order to build and embedded system for door unlocking 10E8 at University of Information Technology - UIT. Further implementation on Raspberry pi 4 will not be updated for privacy concerns. The ONNX model can be used for academic and reproduction purposes only, but any commercial use is prohibited.