Lecture 18: Hough Transform

[@ahme0307]

Thank you Prof. Mubarek Shah (IEEE fellow). I have tried to follow your lecture and implement like Harris IP detector, optical flow, Fundamental matrix, meanshift tracker, Hough transform for line detector. This are great lectures for self learners like me.

[@TheDionator]

It would be better if we could see some of the things he shows on the other screen which we don't see.

[@erickim555]

Great lecture -- very clear and helpful explanation of the Hough Transform (especially the line/circle fitting cases). For the first 2D line-fitting case, I liked his approach of first starting in the slope-intercept space (m-c space), and then moving to the theta-p space to deal with lines with infinite slope (vertical lines). Personally, I could visualize the x-y -> slope-intercept space better.

[@Rohit-oz1or]

Really wonderful. Clearly explained

[@ragzneo1]

Thanks a lot Prof. Mubarek Shah. Clear explanation.

[@hasnainrashid7038]

what he was doing at 18:30 ..this one this one this one and this... seems like he is reading from somewhere at least he should have used a mouse cursor for better understanding

[@almogdavid]

Great videos, thanks

[@tauseefahmad7697]

Thanks sir very explained lecture long live prof

[@ashutoshsingh544]

cant we get two values of m and c with same probability...what to do in that case?

[@fablungo]

I imagine that if they are close, then it might be noise - take the average; if they are far apart then you probably have 2 lines.

[@najwaahmed6503]

awesome video.. thanks