You only look once
计算MAP
https://www.zhihu.com/question/53405779
http://tarangshah.com/blog/2018-01-27/what-is-map-understanding-the-statistic-of-choice-for-comparing-object-detection-models/
http://homepages.inf.ed.ac.uk/ckiw/postscript/ijcv_voc09.pdf
http://host.robots.ox.ac.uk/pascal/VOC/voc2012/htmldoc/devkit_doc.html#sec:ap
code
https://github.com/dmlc/gluon-cv/blob/master/gluoncv/utils/metrics/voc_detection.py
Batch Normalization
https://arxiv.org/pdf/1502.03167v1.pdf
top1 top5
[...] where the top-5 error rate is the fraction of test images for which the correct label is not among the five labels considered most probable by the mode.
First, you make a prediction using the CNN and obtain the predicted class multinomial distribution (∑pclass=1∑pclass=1).
Now, in the case of top-1 score, you check if the top class (the one having the highest probability) is the same as the target label.
In the case of top-5 score, you check if the target label is one of your top 5 predictions (the 5 ones with the highest probabilities).
In both cases, the top score is computed as the times a predicted label matched the target label, divided by the number of data-points evaluated.
Finally, when 5-CNNs are used, you first average their predictions and follow the same procedure for calculating the top-1 and top-5 scores.
top1-----就是你预测的label取最后概率向量里面最大的那一个作为预测结果,如过你的预测结果中概率最大的那个分类正确,则预测正确。否则预测错误
top5-----就是最后概率向量最大的前五名中,只要出现了正确概率即为预测正确。否则预测错误。
Multi-scale training
https://arxiv.org/pdf/1805.09300.pdf
https://stackoverflow.com/questions/50005852/perform-multi-scale-training-yolov2
YOLOv2
https://blog.csdn.net/u010167269/article/details/52638771
https://blog.csdn.net/jesse_mx/article/details/53925356
随机推荐
- git同步遇到报错“fatal: unable to access 'https://github.com/lizhong24/mysite2.git/': Peer reports incompatible or unsupported protocol version.”
git同步遇到报错“fatal: unable to access 'https://github.com/lizhong24/mysite2.git/': Peer reports incompat ...
- java基础03变量和基本数据类型
package cn.bdqn.test; /** * * @author 小豆腐 * * 变量:会变化的量?? * 一个数据在内存中存储空间的表示!在运行期间可以动态改变! * * 关键字:在jav ...
- LRU算法的Python实现
http://flychao88.iteye.com/blog/1977653文章中介绍了常见的几种缓存淘汰策略 LRU:least recently used,最近最少使用算法.其实就是按使用时间倒 ...
- var 声明变量的变量提升问题(let/const)
在ES6之前,JavaScript没有块级作用域(一对花括号{}即为一个块级作用域),只有全局作用域和函数作用域.变量提升即将变量声明提升到它所在作用域的最开始的部分.既全局变量. <scrip ...
- linux使用nohup命令后台运行程序
在linux服务器上搭建web服务器,用ssh客户端登陆后使用./startservice.sh脚本启动服务,但是当ssh断开连接后起的服务也会停掉. 这时可以用nohup ./startservic ...
- select 自动选择 检查下拉列表
下面我们来看一下selenium webdriver是如何来处理select下拉框的,以Apple注册页面为例. https://appleid.apple.com/cgi-bin/WebObject ...
- android 读取通讯录显示到gridview
........... <GridView android:id="@+id/gridView1" android:layout_width="match_pare ...
- [笔记] 基于nvidia/cuda的深度学习基础镜像构建流程
基于NVidia开源的nvidia/cuda image,构建适用于DeepLearning的基础image. 思路就是先把常用的东西都塞进去,再装某个框架就省事儿了. 为了体验重装系统的乐趣,所以采 ...
- DevOps架构实践
1. 场景 持续部署:业界没有统一明确地定义,简单理解为将集成结果部署到不同的环境供用户使用,并且立即反馈部署结果的实践,其中不同的环境包括:开发环境.测试环境.预发布环境.生产环境 持续部署两个核心 ...
- 跨域问题-nginx代理
有时候我们跨域是借助了浏览器对 Access-Control-Allow-Origin 的支持.但有些浏览器是不支持的,所以这并非是最佳方案现在我们来利用nginx 通过反向代理 满足浏览器的同源策略 ...