Skip to content

In order to take my own model on the libfacedetection frame, add some useful functions like depthwise convolution, average pooling, prelu and so on.

Notifications You must be signed in to change notification settings

Rick51/libfacedetection_learning

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

libfacedetection_learning

近段时间,为了将自己的模型适配到于老师的libfacedetection框架上,就阅读了下老师的源码,受益良多。并且因为自己的模型有些特殊层,所以添加了些实现,比如depthwise conv, average pooling, prelu 等等,记录下来,一起学习。

1. Processing

raw_Input:

image

padding:

image

scaleing:(padding黑边的基础上放缩到320X240):

image

output:

image

face[0]: x1: 225, y1: 43, x2: 304, y2: 122

face[1]: x1: 268, y1: 387, x2: 332, y2: 451

face[2]: x1: 364, y1: 281, x2: 447, y2: 364

face[3]: x1: 218, y1: 293, x2: 292, y2: 368

face[4]: x1: 141, y1: 37, x2: 215, y2: 112

2. How to use:

cmake .

make

python jpg2bin.py images/forever2.jpg (因为main.cpp里面的输入是bin文件而不是利用opencv库读取图片,所以先把jpg转bin)

./facedetect_demo 500 500 (按顺序输入图片的实际宽,高)

3. blog记录

关于libfacedetection的一些思考

About

In order to take my own model on the libfacedetection frame, add some useful functions like depthwise convolution, average pooling, prelu and so on.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages