Sign in to follow this  
jenny_wui

Need some suggestion about image processing

Recommended Posts

jenny_wui    200

Hello, I need some suggestion about image processing to know whether it can be done using C /c++ as I am not familiar with MATLAB. I have attached one image. The arrow in the image show the indented circlular portion. I need to crop out only the portion inside the indented portion. I think there should be a sharp difference in gradient intensity in pixel at the indented portion. Is it possible to extract that portion out by reading the binary image?

Thanks in advance.

Share this post


Link to post
Share on other sites
L. Spiro    25638

Most likely the answer is Yes, and it is likely very simple to do so, especially if you know the format of the image and its pixels are fixed-width.

 

However, before anyone can give you a conclusive answer, you should attach whatever image it was that you had planned to attach.

 

 

L. Spiro

Share this post


Link to post
Share on other sites
jenny_wui    200

I shall work with image which is very similar to the one attached and all the images have very similar topology. But I am a bit confused, do I only need to compare pixel values  to extract that porition out or I need to do some more thing? Thanks.

Share this post


Link to post
Share on other sites
Khatharr    8812

Are you asking about a general solution or a specific solution? If general, how similar are other images of the type? Contrast-based edge-finding here is probably going to return a false positive for that upper ridge. I don't have a huge amount of experience (any) with that kind of technique, but I'd wager that this kind of image is a difficult one. I'm interested to hear from graphics gurus about this.

Share this post


Link to post
Share on other sites
redw0lf    334

I've done something similar to this for smaller dots in an microscope image. What i've done first is using a graphic programm like gimp or photoshop to change the image to a full greyscale image, then toying around with the contrast, brightness and edge detection filter settings. After I have found a setting or a procedure which suited my needs I tried to recreate the process in a programming language of my choice (it was. Matlab, R, java or python don't really remember) afterwards applying a cluster algorithm in my case dbscan http://en.wikipedia.org/wiki/DBSCAN to find the different "holes" in a membrane in my case.

If you do it right you may receive something like an O shaped cluster, where you can caculate the center and the radius to one of the outer pixels, which defines you the circle you may try to find.

That would be the way I would try to solve this problem

Share this post


Link to post
Share on other sites
jenny_wui    200

The images are more or less similar.  I have done some thresholding first. I have attached that picture after thresholding. Now the boundary is more prominent with white outline. Now I would like to to move from the center towards the radius to locate that boundary. I need some suggestion about how I can I locate the boundary by comparing the neighboring pixels. Please provide me some suggestions. It would be good for me if I can do the work using C/ C++. Thanks for your suggestions.

Share this post


Link to post
Share on other sites
Khatharr    8812

I have a vague notion that if you could estimate the center point then sample the total value of increasingly larger circles then you could find the ring when the contrast changes. I've never studied computer vision, though, and I can see a lot of potential problems with that method. What are these pictures of, by the way?

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Sign in to follow this