Automatic initialization of active contour model in Medical image segmentation
Give some algorithm or key points for this to do my research
As active contour is a local segmentation technique, this is a very difficult problem in general. Many specific problems rely on the anatomic and/or imaging modality specificity to help in "automatic" initialization of active contours. You can see what they do in the following papers:
Article Automatic Initialization Active Contour Model for the Segmen...
Article Automatic Initialization of Active Contours and Level Set Me...
Conference Paper Automatic initialization for active contour model in breast ...
Good luck!
Active contours or snakes are used interactively: they 'correct' the user when the user is deviating 'too much' from the high gradient, or when the user is making 'too sharp' angles. A starting point will not help you much.
You might give the snake an entire contour as input: it will smooth the sharp angles and stay closer to max-gradient. But ask yourself: why would the snake be better than the original contour it has been fed?
There are many algorithms that can help you in your proposed research, I will suggest you to go through Chan-Vese model for image segmentation, and then try the others.
- Badges
- Science topic
- Similar topics
- Computer Science
- Images