Object Detection论文笔记(4)
来源:互联网 发布:电梯运行优化 编辑:程序博客网 时间:2024/05/17 03:29
Faster R-CNN: Towards real-time object detection with region proposal networks
说明: introduce a Region Proposal Network (RPN) that shares full-image convolutional features with the detection network, thus enabling nearly cost-free region proposals。An RPN is a fully-convolutional network that simultaneously predicts object bounds and objectness scores at each position.
Introduction
1,region proposals are the computational bottleneck in state-of-the-art detection systems。
Region Proposal Networks
1,A Region Proposal Network (RPN) takes an image (of any size) as input and outputs a set of rectangular object proposals, each with an objectness score。
2,This network is fully connected to an n × n spatial window of the input conv feature map. Each sliding window is mapped to a lower-dimensional vector (256-d for ZF and 512-d for VGG). This vector is fed into two sibling fully-connected layers—a box-regression layer (reg) and a box-classification layer (cls)。
3,predict k region proposals, so the reg layer,has 4k outputs encoding the coordinates of k boxes。
A Loss Function for Learning Region Proposals
assign a positive label to two kinds of anchors: (i) the anchor/anchors with the highest Intersectionover-Union (IoU) overlap with a ground-truth box, or (ii) an anchor that has an IoU overlap higher than 0.7 with any ground-truth box.
Optimization
1,The RPN, which is naturally implemented as a fully-convolutional network, can be trained end-to-end by back-propagation and stochastic gradient descent (SGD) 。
2,we randomly sample 256 anchors in an image to compute the loss function of a mini-batch, where the sampled positive and negative anchors have a ratio of up to 1:1。
3,all new layers的权值初始化:高斯分布(μ=0,σ=0.01),all other layers(比如共享卷积层)用ImageNet来权值初始化。用ZF net来进行进行微调。学习率:0.001(60k)->0.0001(20k)。动量:0.9。weight decay: 0.0005。
Sharing Convolutional Features for Region Proposal and Object Detection
1,sharing convolutional layers between the two networks, rather than learning two separate networks
2, 4-step training algorithm to learn shared features via alternating optimization
Implementation Details
1,Multi-scale与speed-accuracy之间的trade-off
2,To reduce redundancy, we adopt non-maximum suppression (NMS) on the proposal regions based on their cls scores.
Faster R-CNN
- Object Detection论文笔记(4)
- Object Detection论文笔记(1)
- Object Detection论文笔记(2)
- Object Detection论文笔记(3)
- Object Detection 系列论文笔记
- [论文笔记4]Robust Object Detection via soft cascade
- 【object detection】R-cnn论文笔记
- Object Detection论文清单
- [深度学习论文笔记][Object Detection] Rich feature hierarchies for accurate object detection and semantic seg
- [深度学习论文笔记][Object Detection] You Only Look Once: Unified, Real-Time Object Detection
- 论文笔记(1)DenseBox: Unifying Landmark Localization with End to End Object Detection
- 论文笔记(3)You Only Look Once:Unified, Real-Time Object Detection
- 目标检测 Feature Pyramid Networks for Object Detection(FPN)论文笔记
- 【论文笔记】Scalable Object Detection using Deep Neural Networks
- 【深度学习论文笔记】Deep Neural Networks for Object Detection
- 论文阅读笔记:Object Detection Networks on Convolutional Feature Maps
- 论文笔记 《Deep Neural Networks for Object Detection》
- 【论文学习笔记】Class-Specific Hough Forests For Object Detection
- 一篇文章讲透CDN HTTPS安全加速基本概念、解决方案及优化实践
- 16秋计算机JAVA第三节课作业(钟永钜) 4~6题
- Java的Json解析包FastJson使用
- CSS中定义变量,并使用变量设置属性值
- Python模块_PyLibTiff读取tif文件
- Object Detection论文笔记(4)
- 顺序堆栈(数组)
- 从数据库得到数据导出指定格式的xml文件,上传到NC接口,返回回执到本地一个xml文件
- 栈的压入、弹出序列
- Inno Setup入门(五)——添加readme文件
- JAVA_int类型数据精度高于float低于double
- 下面总结 8 组常用的Eclipse快捷键
- Java发送邮件的简单实现
- kickstart无人值守安装