https://github.com/zdfb/Yolov5_face

yolov5人脸检测，基于cspdarknet_s的wider_face性能为:(1)easy:94.58%；(2)medium：93.37；（3）82.84%
https://github.com/zdfb/Yolov5_face

Last synced: 2 months ago
JSON representation

yolov5人脸检测，基于cspdarknet_s的wider_face性能为:(1)easy:94.58%；(2)medium：93.37；（3）82.84%

• Host: GitHub
• URL: https://github.com/zdfb/Yolov5_face
• Owner: zdfb
• Created: 2022-06-06T08:17:03.000Z (over 2 years ago)
• Default Branch: main
• Last Pushed: 2022-06-07T03:07:42.000Z (over 2 years ago)
• Last Synced: 2024-02-11T22:47:27.574Z (11 months ago)
• Language: Python
• Homepage:
• Size: 3.99 MB
• Stars: 5
• Watchers: 1
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

• awesome-face-detection-and-recognition - zdfb/Yolov5_face
• awesome-yolo-object-detection - zdfb/Yolov5_face
• awesome-yolo-object-detection - zdfb/Yolov5_face

README

        
# YOLOV5_Face

基于pytorch的Yolov5人脸检测

## 性能效果

| Backbone| easy | medium | hard |

|:-|:-:|:-:|:-:|

| CSPdarknet-s | 94.58% | 93.37% | 82.84% |

## 预训练模型

+ 基于cspdarknet_s的人脸检测模型yolov5_face。


>- 链接: https://pan.baidu.com/s/1Gm4m7DNbSmmbyQ43NZHojA 

>- 提取码：il9n

## 训练

### 1. 下载数据集并放置在Data下

>- 链接: https://drive.google.com/open?id=11UGV3nbVv1x9IC--_tK3Uxf7hA6rlbsS

### 2. 开始训练

``` bash

python train.py

```

## 测试图片

修改utils/utils_yoloface.py文件中的model_path指向训练好的模型。

在predict.py文件下输入图片路径，运行：

``` bash

python predict.py

```

## widerface 数据集测试

修改utils/utils_yoloface.py文件中的model_path指向训练好的模型。

``` bash

python test_widerface.py

```

``` bash

cd ./widerface_evaluate

python setup.py build_ext --inplace

python evaluation.py

```

## 人脸检测示意图




## Reference

- https://github.com/biubug6/Pytorch_Retinaface

- https://github.com/bubbliiiing/yolov5-pytorch