face recognition based on yolov8 & Arcface

author: HBen
Edit Time：2024.02.26

Structure

Yolov8_Arcface
│   README.md
│   main.ipynb:testing process  
│   main.py:Use capture of the computer get the image . then show the detection image 
│   Add_face_info.py: trans the face to feature ,then store to datasets. 
|   config.json: project define
|   face_dataset.json: face datasets
│
└───model
│   │   YOLOV8.py: face Detection
│   │   Arcface.py: face recognition
│      
└───weights
│   │   yolov8n-face.onnx
│   │   Arcface_rs50.pth
│   
└───util
│   │   facial_pro.py : store some function that process image(alignment、enhancement)
│   │   matlab_cp2form.py:store some function that calculate the similarity
│     
└───ADD_IMAGE
│   │   1.jpg
│   │   ……
└───

How to use

• environments:
  • python >= 3.7
  • opencv-python >=4.7.0.72
  • pytorch >= 1.12.0
• Download weight
  • Arcface_rs50.pth（链接：https://pan.baidu.com/s/1dna2U7CwUg2Jq6-b_JV4aw?pwd=s2f1 提取码：s2f1）
  • yolov8n-face.onnx （链接：https://pan.baidu.com/s/1jCPZcQhFIZZzj1UOOrLOdw?pwd=s2f1 提取码：s2f1）
• Run
  • Add face image to ADD_IMAGE floder and set the face_id as the image name eg: 1.jpg（the guy who in the image id is 1）
  • Automatically convert the face image under the ADD_IMAGE to save the face-dataset.json file as facial features

      python Add_face_info.py
  

  • Then start use

      python main.py