Recognition of objects in almost plain background

1

I want to recognize climbing holds in a climbing wall. I thought about implementing my own heuristic algorithm for background detection because it's almost plain, but I was wondering if this kind of task is well known in the community.

Does anybody know the right keywords associated with this? Any research papers or tools that are relevant?

I'm mostly interested in non-supervised learning to avoid having to create a training set for what I think it should be relatively easy but I might be very wrong.

This is a sample image I'd like to classify: enter image description here You can see 4 big holds, 2 small and a number of holes. Ideally I'd detect all of those with bounding boxes but having only the big holds would be good enough. Any hint on how to approach this would be highly appreciated! Thanks

Quimey

Posted 2019-01-18T01:59:16.543

Reputation: 111

2Try salient object detection. – Mark.F – 2019-01-18T06:12:15.463

1

@mark It works great. Using this https://github.com/Joker316701882/Salient-Object-Detection I've got https://imgur.com/a/TMHt7lx

– Quimey – 2019-01-18T11:54:17.127

No answers