Best Segmentation technique for a structurally homogenous image

I would like to segment a structurally homogenous object from the background image. Techniques like active contours, region growing and merging and adaptive thresholding have already been employed and little efficiency has been demonstrated. The image is a gray scale image and i would like to make use of a new technique to segment these structurally homogenous objects from the background. Kindly suggest me a better segmentation technique other than the above mentioned techniques.

[Image Analyst]

In order to give image processing advice, we always prefer it if you post an image. Otherwise we're operating blindly. Sorry, but we're funny that way.

[shivasakthi]

[shivasakthi]

I would like to segment the structurally homogenous object on the right from the above image. Techniques like active contours, region growing and merging and adaptive thresholding have already been employed and little efficiency has been demonstrated. The image is a gray scale image and i would like to make use of a new technique to segment these structurally homogenous objects from the background. Kindly suggest me a better segmentation technique other than the above mentioned techniques. Also help me with a matlab code for the same.

[Image Analyst]

Why can't you simply use size and/or shape information? It's big and round so that should be a major clue. Did you see my BlobsDemo in my File Exchange?

Answer 1

[Spandan Tiwari]

Try using the new function activecontour in the Image Processing Toolbox that shipped in R2013a. It should be useful for this problem.