一、 背景介绍
- 三种强大的物体识别算法——SIFT/SURF、haar特征、广义hough变换的特性对比分析
https://blog.csdn.net/cy513/article/details/4285579 - 图像处理之Haar特征
https://blog.csdn.net/u012507022/article/details/54138299 - 哈尔小波变换的原理及其实现(Haar)
https://blog.csdn.net/victory06057231/article/details/6708330 - DPM(Deformable Part Model)原理详解
https://blog.csdn.net/qq_14845119/article/details/52625426 - DPM(Deformable Parts Models)-----目标检测算法理解
https://blog.csdn.net/qq_22625309/article/details/72493223 - HOG特征(Histogram of Gradient)学习总结
https://www.cnblogs.com/wyuzl/p/6792216.html- 直方图(histogram)中的bins应如何理解
https://blog.csdn.net/u011776903/article/details/74637080 - 多尺度与多分辨率的理解
https://blog.csdn.net/u013360881/article/details/43452057 - 什么是图像金字塔
- https://www.jianshu.com/p/cefbdc2dc5b9*
- 图像金字塔总结
https://blog.csdn.net/xiamentingtao/article/details/78596240
- 直方图(histogram)中的bins应如何理解
- 尺度不变特征变换匹配算法详解(暂未阅读)
https://blog.csdn.net/chezhai/article/details/66044054?utm_source=copy - 提高目标检测精度的常用方法(含部分代码)
https://zhuanlan.zhihu.com/p/42946589 - NMS——非极大值抑制
https://blog.csdn.net/shuzfan/article/details/52711706 - 非极大值抑制(Non-Maximum Suppression,NMS)
https://www.cnblogs.com/makefile/p/nms.html
代码实现:
https://blog.csdn.net/williamyi96/article/details/77996167
https://www.cnblogs.com/makefile/p/nms.html - 目标检测中region proposal的作用?
https://www.zhihu.com/question/265345106
二、YOLO
- 目标检测|YOLO原理与实现
https://zhuanlan.zhihu.com/p/32525231 - YOLO
https://blog.csdn.net/guoyunfei20/article/details/78744753 - [目标检测]YOLO原理
https://www.cnblogs.com/fariver/p/7446921.html - YOLO配置文件理解
https://blog.csdn.net/hrsstudy/article/details/65447947 - Yolo-实时目标检测算法训练自己的数据集教程
https://www.jianshu.com/p/c1a009385f59 - yolo的训练和测试
https://blog.csdn.net/qq_30401249/article/details/51564871