Edge detection with prior shapes based on Mumford-Shah modelPP

Edge detection plays an immensely important role in image processing. In this paper, we propose a new model with the prior shapes on the basis of the well-known Mumford Shah model. To solve this minimum model, we design an efficient algorithm combined with the fixed-point iterative method, the Split-Bregman (SB) method and the Lasso method. Experimental results show that the proposed model and algorithm can get better detected edges and have more advantages in efficiency and accuracy for different pure and noisy images, even for several different shapes.

This is poster number 26 in Poster Session

Yuying Shi (North China Electric Power University)
Yilin Li (North China Electric Power University)
Juan Zhang (North China Electric Power University)
image segmentation, inverse problems, partial differential equation models