多阶段细化分割-iccv2017-A Stagewise Refinement Model for Detecting Salient Objects in Images
来源:互联网 发布:怎么样打造淘宝爆款. 编辑:程序博客网 时间:2024/05/22 15:50
读论文《A Stagewise Refinement Model for Detecting Salient Objects in Images》,
论文源码已经公布:https://github.com/TiantianWang/ICCV17_SRM
里面提到了PPM,
查PPM的实现,找到论文《Pyramid Scene Parsing Network》(CVPR2018),基本上与这篇论文一致,只是最后的concat不一样。
PSPnet论文有caffe实现的源码,地址在:https://github.com/hszhao/PSPNet
把里面的ppm实现和双插值的prototxt贴在这,以备以后使用。
layer {
name: "conv5_3_pool1"
type: "Pooling"
bottom: "conv5_3"
top: "conv5_3_pool1"
pooling_param {
pool: AVE
kernel_size: 60
stride: 60
}
}
类似的kernel_size 和stride 改成30,20,10。共四个平均池化,生成1*1,2*2,3*3,6*6的池化bin。
之后的双线性插值上采样是用了一个新的层Interp。
layer {
name: "conv5_3_pool1_interp"
type: "Interp"
bottom: "conv5_3_pool1_conv"
top: "conv5_3_pool1_interp"
interp_param {
height: 60
width: 60
}
caffe.proto中的参数设置:
message InterpParameter {
optional int32 height = 1 [default = 0]; // Height of output
optional int32 width = 2 [default = 0]; // Width of output
optional int32 zoom_factor = 3 [default = 1]; // zoom factor
optional int32 shrink_factor = 4 [default = 1]; // shrink factor
optional int32 pad_beg = 5 [default = 0]; // padding at begin of input
optional int32 pad_end = 6 [default = 0]; // padding at end of input
}
PSPnet中提到的中间损失auxiliary loss的描述,只在训练阶段用,测试时去掉,因此在caffe的源码中没有找到。
在知乎中有关于这篇论文的讨论,https://www.zhihu.com/question/53356671
在说这个额外损失,有人解释说:只是额外一个正常的softmax loss
Apart from the main branchusing softmax loss to train the final classifier, another classifieris applied after the fourth stage, i.e., the res4b22residue block. Different from relay backpropagation that blocks the backward auxiliary loss to several shallowlayers, we let the two loss functions pass through all previouslayers. The auxiliary loss helps optimize the learningprocess, while the master branch loss takes the most responsibility.We add weight to balance the auxiliary loss.
Setting an appropriate loss weight α in the auxiliarybranch is important. ‘AL’ denotes the auxiliary loss. Baseline isResNet50-based FCN with dilated network. Empirically, α = 0.4 yields the best performance.
- 多阶段细化分割-iccv2017-A Stagewise Refinement Model for Detecting Salient Objects in Images
- Detecting Faces in Images: A Survey( 翻译)
- Open Sourcing a Deep Learning Solution for Detecting NSFW Images
- Thinking about a paper "A Refinement Approach to Handling Model Misfit in Text Categorization"
- Interactive Graph Cuts for Optimal Boundary and Region Segmentation of Objects in N-D ImageS
- Learning to Count Objects in Images
- Detecting a mobile browser in ASP.NET
- Detecting a mobile browser in ASP.NET
- 《Interactive graph cuts for optimal boundary & region segmentation of objects in N-D images》读后感及问题
- 图割论文阅读笔记:Interactive Graph Cuts for Optimal Boundary & Region Segmentation of Objects in N-D Images
- 物体计数--Learning To Count Objects in Images
- Rethinking "A refinement..."
- SegLink on github-Detecting Oriented Text in Natural Images by Linking Segments
- Detecting Skin in Images & Video Using Python and OpenCV皮肤检测
- 语义分割--RefineNet: Multi-Path Refinement Networks for High-Resolution Semantic Segmentation
- 语义分割--Label Refinement Network for Coarse-to-Fine Semantic Segmentation
- 图像分割“RefineNet-Multi-Path Refinement Networks for High-Resolution Semantic Segmentation”
- 论文笔记:A Global Covariance Descriptor for Nuclear Atypia Scoring in Breast Histopathology Images
- Raid5两块硬盘掉线怎么恢复数据_北亚案例
- 浅谈web前端安全
- 泛型
- [牛客网]字符串价值
- 神奇的 routing mesh
- 多阶段细化分割-iccv2017-A Stagewise Refinement Model for Detecting Salient Objects in Images
- 一元三次方程求解
- c#动态生成控件,并为控件注册事件
- Spring的两种代理JDK和CGLIB的区别浅谈
- 对ios11 适配问题1——tableView的下拉刷新控件下移
- Kafka 设计与原理详解
- SpringBoot 使用logback日志分级别生成文件
- 关于ant-Design的一些问题
- 一道关于JavaScript解析器错误的面试题