| Paper: | IMDSP-P11.1 | ||
| Session: | Image Segmentation | ||
| Time: | Thursday, May 18, 14:00 - 16:00 | ||
| Presentation: | Poster | ||
| Topic: | Image and Multidimensional Signal Processing: Image Segmentation | ||
| Title: | MUMFORD-SHAH MODEL WITH FAST ALGORITHM ON LATTICE | ||
| Authors: | Lu Yu, Qiao Wang, Lenan Wu, Southeast University, China; Jun Xie, PLA University of Science and Technology, China | ||
| Abstract: | Mumford-Shah piecewise smooth functional is a variational PDE model widely used in image segmentation and smoothing. An analogous discrete model which models image as an MRF has also been built. In this paper, we propose another discrete Mumford-Shah piecewise smooth model on lattice from a different perspective, as well as a method to find the solution. Only two simple and deterministic optimization techniques are used in the model to seek the solution. Compared with traditional continuous model and the existing discrete model, the model in this paper is much simpler and the approach is much easier and faster. Furthermore, experiment results show that the model performs better on various aspects. | ||
©2018 Conference Management Services, Inc. -||- email: webmaster@icassp2006.org -||- Last updated Friday, August 17, 2012