论文阅读:Prominent Object Detection and Recognition: A Saliency-based Pipeline
论文阅读:Prominent Object Detection and Recognition: A Saliency-based Pipeline

如上图所示,本文旨在解决一个问题:给定一张图像,我们最应该关注哪些区域?怎么将其分割出来?这是一个什么东东?这三个子问题为一体。
Problem formulation: Given an image, determine the most influential item in the scene in terms of region of interest, pixel-level extent (segmentation), and object type.
作者提出的框架为:

可以发现,上述流程图就是各个需要完成任务的级联和组合。
网络的训练也是各个级别依次完成的。没啥好说的。
有些实验结果看起来还是挺烂的:

有的看起来还可以,但是也不怎么好:

论文阅读:Prominent Object Detection and Recognition: A Saliency-based Pipeline的更多相关文章
- Object Detection with Discriminatively Trained Part Based Models
P. Felzenszwalb, R. Girshick, D. McAllester, D. RamananObject Detection with Discriminatively Traine ...
- [论文阅读] Joint Face Detection and Alignment using Multi-task Cascaded Convolutional Networks(MTCNN)
相关论文:Joint Face Detection and Alignment using Multi-task Cascaded Convolutional Networks 概论 用于人脸检测和对 ...
- 读论文系列:Object Detection CVPR2016 YOLO
CVPR2016: You Only Look Once:Unified, Real-Time Object Detection 转载请注明作者:梦里茶 YOLO,You Only Look Once ...
- 读论文系列:Object Detection ECCV2016 SSD
转载请注明作者:梦里茶 Single Shot MultiBox Detector Introduction 一句话概括:SSD就是关于类别的多尺度RPN网络 基本思路: 基础网络后接多层featur ...
- 读论文系列:Object Detection SPP-net
本文为您解读SPP-net: Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition Motivat ...
- 读论文系列:Object Detection NIPS2015 Faster RCNN
转载请注明作者:梦里茶 Faster RCNN在Fast RCNN上更进一步,将Region Proposal也用神经网络来做,如果说Fast RCNN的最大贡献是ROI pooling layer和 ...
- 读论文系列:Object Detection ICCV2015 Fast RCNN
Fast RCNN是对RCNN的性能优化版本,在VGG16上,Fast R-CNN训练速度是RCNN的9倍, 测试速度是RCNN213倍:训练速度是SPP-net的3倍,测试速度是SPP-net的3倍 ...
- 目标检测(一)RCNN--Rich feature hierarchies for accurate object detection and semantic segmentation(v5)
作者:Ross Girshick,Jeff Donahue,Trevor Darrell,Jitendra Malik 该论文提出了一种简单且可扩展的检测算法,在VOC2012数据集上取得的mAP比当 ...
- 【计算机视觉】Selective Search for Object Recognition论文阅读3
Selective Search for Object Recoginition surgewong@gmail.com http://blog.csdn.net/surgewong 在前 ...
随机推荐
- B树,B+树,B*树以及R树的介绍
https://blog.csdn.net/peterchan88/article/details/52248714 作者:July.weedge.Frankie.编程艺术室出品. 说明:本文从B树开 ...
- Real Time Credit Card Fraud Detection with Apache Spark and Event Streaming
https://mapr.com/blog/real-time-credit-card-fraud-detection-apache-spark-and-event-streaming/ Editor ...
- caffe_ssd学习-用自己的数据做训练
几乎没用过linux操作系统,不懂shell编程,linux下shell+windows下UltraEdit勉勉强强生成了train.txt和val.txt期间各种错误辛酸不表,照着examples/ ...
- python 创建二维数组的方法
废话不多说,直接上代码: #coding=utf-8 def two_di_demo1(): a=[] for i in range(10): a.append([]) for j in range( ...
- python递归的例子
例子1:递归实现嵌套列表求和 #encoding=utf-8 a=[[1,2,3], [4,5,6], [7,8,9]]def listsum(L): result=0 for i i ...
- Linux基础命令---traceroute追踪路由
traceroute traceroute指令输出到目标主机的路由包.Traceroute跟踪从IP网络到给定主机的路由数据包.它利用IP协议的生存时间(TTL)字段,并试图在通往主机的路 ...
- vue 去掉路由中的#
在router.js中修改, const router = new VueRouter({ mode: 'history', routes: [...] })
- The Little Prince-11/30
The Little Prince-11/30 Today, I have a meeting in our department. I sincerely hope that all of my d ...
- Python进阶【第六篇】内置函数中好玩的几个(今天写的太水)
zip()函数 两个参数一一对应,参数是序列类型,序列包括列表,元组,字符串,当两个序列不等长时,按公共最长部分匹配,形似“拉链”. max()和min()函数 以max()为例,min()类似,只是 ...
- P2P原理及UDP穿透简单说明(转)
源: P2P原理及UDP穿透简单说明