Shape Modeling with Front Propagation: A Level Set Approach
IEEE Transactions on Pattern Analysis and Machine Intelligence
Robust 3D Segmentation of Anatomical Structures with Level Sets
MICCAI '00 Proceedings of the Third International Conference on Medical Image Computing and Computer-Assisted Intervention
ICCV '95 Proceedings of the Fifth International Conference on Computer Vision
Hi-index | 0.00 |
A number of methods based on geometric deformable models have been proposed recently to segment medical images. These methods require the definition of a speed function that governs model deformation. In this paper we propose a new speed function that is modified dynamically as the front progresses. This new speed function is particularly well adapted to situations where edges are ill-defined, adjacent structures have comparable intensity values, and images are noisy. We illustrate qualitatively the performance of our approach on a variety of MR, CT, and ultrasound images. The examples show the generality of our approach and its insensitivity to parameters. We also evaluate it quantitatively on several CT scans of the liver.