Active Contours: A New Distribution Metric for Image Segmentation

We present a new distribution metric for image segmentation for active contours.

Description

How can we characterize an image, and does the same characterization yield the same result? In this work, we study one possible characterization, distribution metrics. That is, we assume the desired region of interest has a different probability distribution from its corresponding background. Using this, we present a new distribution metric for image segmentation that arises as a result in prediction theory. Forming a natural geodesic, our metric quantifies “distance” for two density functionals as the standard deviation of the difference between logarithms of those distributions. Using level set methods, we incorporate an energy model based on the metric into the Geometric Active Contour framework. We also demonstrate the algorithm on several challenging medical images, which further ensure the viability of the metric in the context of image segmentation.

For More Information: www.romeilsandhu.com/research_projects/p

• MATLAB release: MATLAB 7.9 (R2009b)
• Other requirements: Everything is coded in MATLAB

[Swati]

Hi Sandhu. I wish to apply your algorithm on a set of images in time sequence with 30 frames per second for 5 seconds. So there are approx 150 images and i wish the code to detect the same contour in all the images. I am successfull in doing this on one image

[Michele]

when I compile the code, I want visualize only segmented region.
However, the code displays the segmented region in black.
Can you help me?

[Gavin]

Note that this code requires the signal processing toolbox.

[Romeil Sandhu]

Hi Amandeep,

What you will need to have your own initalization is simply comment out the line where I pre-load a binary mask. Also, when you call "run.m", do not give it a mask variable.

This will invoke "run.m" to call a user initialization program.

Best,
~rome

[Ashvin Kumar]

 

[amandeep]

sir, i have checked the same code. Its very good.
But i want to apply the same on ultrasound image. Please tell me how to select the mask and how can i change the mask for a particular image

[Eran Ukwatta]