Object Recognition Using Hog Features
可以将文章内容翻译成中文,广告屏蔽插件可能会导致该功能失效(如失效,请关闭广告屏蔽插件后再试): 问题: Hi I'm going to recognize objects on my images Before creating a dataset to train with svm I just cropped single part of my image computed hog descriptors put them into svmdetector and then tried to find it using multiscale detection on my original image. however detection points out center of my image. I am not sure if I'm doing something wrong with hog detection or I just mark wrong place. this is my image this is cropped part of it and this is the result and my code is below. #include "opencv2/highgui/highgui.hpp" #include "opencv2