Skip to content

jinjinhong/DeepFaceProper

 
 

Repository files navigation

DeepFaceProper

0. 效果展示

image image

1. 数据集及工具

2. 运行命令

  • 2.0 执行python data_wash.py筛选有效数据
  • 2.1 执行python train.py训练数据
  • 2.2 执行Python convert.py转换模型
  • 2.3 执行python test.py测试单张图片数据

3. 参考链接

4. 问题

  • 年龄是一个比较感性的问题,对于老人小孩区分可能比较明显,但是对于中年人效果不是很好,数据会抖动厉害
  • 对于同一张人脸,人靠近摄像头程度不同,光照影响的不同,数据都会剧烈变化

5. 更新

About

Using wideRestnet to get proper with dataset of imdb face

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%