Assume that you have an image that you going to classify its class ID.

One good option is to use interest points e.g FAST that can find coners/edges inside an image.

enter image description here

From these points you could classify a specific area....but how do I know which area I should classfy?

Assum that you're having a bunch of read dots over an image.The red dots covering different types of objects, which means that there are multiple objects inside the image. Not only just one object.

You want to "cut" out the coordinates of the image that are specific to one object for the image. How can that be done?

Thake the image above. If you want to classify the windows, you need to know that all red dots round a specific region, is one object. But how can I descripte that in such way?

  • $\begingroup$ 1. I don't understand what your question is. I see three questions in the post. Please pick one question that you want answered. We expect you to ask only one question per post. 2. Please proof-read your post and edit it to fix errors (e.g., "classfy", "read", "Thake"). 3. Regarding "how do I know which area I should classfy?" - I don't know, you tell us - you are the one who is defining the problem you want solved. 4. What is a "red dot"? I don't understand the definition of how you have assigned colors to dots. 5. Have you read about semantic segmentation? $\endgroup$
    – D.W.
    Nov 4, 2023 at 7:24


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.