Object Detection入门

来源:互联网 发布:java连接池 编辑:程序博客网 时间:2024/06/07 02:40

简介:

what is object detection
这里写图片描述
Localization and classification in multiple objects images


学习步骤:

  • Tensorflow Object Detection API(用5分钟的时间可以跑一遍感受一下)
  • rbg大神主页:http://www.rossgirshick.info/
  • Kaiming He大神主页:http://kaiminghe.com/
  • 传统目标检测算法
    代表:DPM
    论文:
    A Discriminatively Trained, Multiscale, Deformable Part Model
    Object Detection with Discriminatively Trained Part Based Models
    (项目和论文地址rbg主页上有)
  • 基于深度学习框架 (重点)
    涉及基础网络知识Alexnet, ZFnet, VGGnet, GoogLeNet, ResNet等
    R-CNN:Rich featurehierarchies for accurate object detection and semantic segmentation(项目和论文地址rbg主页上有)
    SPPnet:Spatial Pyramid Pooling in Deep Convolutional Network for Visual Recognition
    论文地址:https://arxiv.org/pdf/1406.4729v2.pdf
    MultiBox
    Fast R-CNN (项目和论文地址rbg主页上有)
    Faster RCNN (项目和论文地址rbg主页上有)
    YOLO: You Only Look Once:Unified, Real-Time Object Detection
    论文地址:https://arxiv.org/abs/1506.02640
    YOLO9000: Better, Faster, Stronger
    论文地址:https://arxiv.org/abs/1612.08242
    项目主页:https://pjreddie.com/darknet/yolo/
    SSD:SSD: Single Shot MultiBox Detector
    论文地址: http://www.cs.unc.edu/~wliu/papers/ssd.pdf
    项目地址:https://github.com/weiliu89/caffe/tree/ssd

论文查看地址:

https://handong1587.github.io/deep_learning/2015/10/09/object-detection.html