Medical Image Segmentation

The main idea here is to isolate and extract individual components from a medical image. This is an important part of medical imaging; once a shape is found, physicians can measure various quantities, such as the size of tumors and the thickness of heart walls. Tracing the boundary of these shapes by hand is time-consuming. Instead, the goal is an automatic segementation technique which finds the desired regions automatically. This can be done using a combination of fast marching methods and level set methods.

To demonstrate, below is a two-dimensional cross-section of a scan across the chest: the patient's feet are pointed out of the screen, and the heart, liver, and various other features are visible.
(If you have trouble getting this page to load, it is a problem of browser compatabilities. Please go here)

Instructions:

1. Move the mouse into a region and click
2. The Java script executes the Fast Marching Method to segment the desired region.
3. Things to try:
   • Click on individual regions to see the sensitivity of the technique. Notice that the technique often finds boundaries that are not easy to see.
   • Try clicking outside the entire region, and see that the front collapses in on the body.

Main Reference: