Ask Your Question
0

SIFT with very similar objects

asked 2020-04-05 11:21:18 -0500

Hi, I'm computing SIFT algorithm with OpenCV on a kind of store shelf where I have to recognize some products. But since in this shelf I have two products that are very very similar except that for the color and for a writing, the algorithm is failing in the keypoint detection highlighting features in the wrong product.

What are your suggestions in order to avoid this misclassification and in order to find the right product on the shelf?

edit retag flag offensive close merge delete

1 answer

Sort by ยป oldest newest most voted
0

answered 2020-04-06 01:25:06 -0500

berak gravatar image

updated 2020-04-06 02:41:47 -0500

where I have to recognize some products.

please don't try with SIFT feature matching, it was not meant to be used this way, and it does not know about "objects" at all

the algorithm is failing in the keypoint detection

that's expected. wrong tool for the job

What are your suggestions

(re)train / finetune some cnn model, like YOLO, SSD, RCNN, and use it with opencv's dnn::Net

edit flag offensive delete link more
Login/Signup to Answer

Question Tools

1 follower

Stats

Asked: 2020-04-05 11:20:25 -0500

Seen: 41 times

Last updated: Apr 05