Highlight a specific part within an image

I would like to ask how to read a picture, and then distinguish if there is something inside this picture from another picture is also included.

Depending on the attached images , I would like, for example, to insert image 1 and then identify and define the non-smooth part based on an image containing a non-smooth part, I will include it.

The goal is to highlight areas of an image that are rough and smooth in nature.

Knowing that the images of the original sample will differ each time, and the system has to distinguish that area due to the same image of the part.

I hope I was able to convey the idea and I hope to find an answer.

thanks

[DGM]

Define "something inside this picture [say picture A] from another picture [B] is also included".

Does that mean that picture B is a subset of A, or does that merely mean that some feature of B is similar to some feature in A? If the latter, what defines "similar"?

If the goal is simply to find the location of textured regions of A, what is the purpose of using B?

[Mariam Ramzi]

DGM

thank you for your comment

Yes the idea is to find specific areas in A (areas of a rough nature) I used B because I could not find another alternative and I thought this is the easiest way, I hope to find a more effective way

Answer 1

[Image Analyst]

Try something like stdfilt() or entropyfilt() to identify the rough region.

[Mariam Ramzi]

Thank you for the answer

I tried it but I didn't get the desired result

[Image Analyst]

Wow, that's really strange. It worked perfectly for me. What did you do to not make it work? If you did what I told you, it would have worked perfectly. While you're posting your non-working code, I'll attach my code so you can see how it should work. See in the display below how it found the rough area and outlined it in red.

[Mariam Ramzi]

Thank you very much for the answer

I appreciate that

But I have a question about the " mask", what exactly does it do, here was my mistake in my previous code

[Image Analyst]

@Mariam Ramzi, I did not see your previous code because you did not attach it. In my code, mask is a binary image gotten by thresholding the standard deviation image. The resulting image is true wherever there is high standard deviation (meaning the image is rough there), and false where there is low standard deviation (meaning the image is smooth there). I can analyze your code for the problem but you'd have to attach it.