CODEWITHSUNDEEP
c#image-processingmachine-learningcomputer-vision
Mircea Paul Muresan
Mircea Paul Muresan

Reputation: 639

Sliding Window Object detection

I am trying to implement a sliding window technique to identify my hypothesis. I am using a 64 x 128 window however I also want to find objects smaller than my window or objects larger than my window. I heard of an apporoach of resizing the image after a full traversal with my kernel window and placing my sliding window at the coordinates obtained from my current possition and the scalling factor, I have no idea how that technique was called... Does anyone know how to solve this issue?

Thank you for your time !

Upvotes: 0

Views: 307

Answers (1)

dajuric
dajuric

Reputation: 2507

I assume that you are referring to pyramidal approach which is frequently used in optical flow. The patch is first detected at upper pyramid scale, than the obtained coordinates are rescaled and the patch is searched again in lower pyramid scale.

Haar object detection does not need image rescaling as the features can be easily rescaled using integral image.

Upvotes: 1

Related Questions