CODEWITHSUNDEEP
opencvmachine-learningcomputer-vision
sub_o
sub_o

Reputation: 2802

Applying mean shift to detected windows

currently I've implemented an algorithm to extract features out of depth images, and I want to apply mean shift clustering on all the detected windows.

How do I use openCV for that ? I understand that OpenCV has mean shift, however it seems like it's applicable on back projection of an image. Is there any other OpenCV functions that I can use to find the mode of the detection windows (maybe utilising something similar to mean shift)

Or is there any open-source mean shift codes that I can repurpose for this matter ?

Upvotes: 1

Views: 843

Answers (1)

Don Reba
Don Reba

Reputation: 14041

OpenCV exposes a mean shift image filter, but not the general clustering algorithm. You can adapt a kernel density estimation library, such as LibAgf, for this purpose fairly easily. Mean shift is simply repeated assignment to every point of the average of all others weighted by KDE.

Upvotes: 1

Related Questions