I have already done human detection but don't know how to crop detected frames of video and save it in video format... here is the code link which i followed https://github.com/ambakick/Person-De...
Plz suggest me !!! Thanks
the main problem here is, that a writing video file needs a constant frame size, while size of your crops will vary due to person size & distance. so try following approach:
Size dstSize(64,128); // dst video size
VideoWriter vid("my.avi", VideoWriter::fourcc('M','J','P','G'), 10, dstSize);
// later, in a loop, detect humans
Rect rec = ...
Mat cropped = frame(rec);
Mat dst;
resize(cropped, dst, dstSize);
vid.write(dst);
Asked: 2020-03-25 16:06:33 -0600
Seen: 1,511 times
Last updated: Mar 26 '20
Surprisily, when you google "opencv write video", you'll find lots of information
Thanks for your reply & I will be definitely search your suggestion on google but i want to crop human detected frames as i already mention above and provide link also( which i followed) On different websites(like stack overflow , git hub, pyImageSearch) they all are talking about the images not video & they all crop image not video frames... I am beginner in Computer vision and i don't have any idea about it... Please help me!!! Thanks!