Normalization on conv4_3 in SSD

版权声明：本文为博主原创文章，转载请注明出处. https://blog.csdn.net/weixin_35653315/article/details/72715367

还是在SSD中发现的：

layer {
  name: "conv4_3_norm"
  type: "Normalize"
  bottom: "conv4_3"
  top: "conv4_3_norm"
  norm_param {
    across_spatial: false
    scale_filler {
      type: "constant"
      value: 20.0
    }
    channel_shared: false
  }
}
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14

有两个问题：
1. Normalize层做了什么？
2. 为什么只对conv4_3进行normalize 操作？

Normalize Layer

Normalize 操作很简单：给定一个向量xx维向量上操作。

为什么只对conv4_3进行norm操作？

https://github.com/weiliu89/caffe/issues/241

That was discovered in my other paper (ParseNet) that conv4_3 has different scale from other layers. That is why I add L2 normalization for conv4_3 only.

References

• https://github.com/weiliu89/caffe/issues/241
• http://www.jianshu.com/p/db35f7f8a593
• ParseNet