Active Contouring With Level Set Regularization
We present an edge-based active contour model, which is an extension of our MAC model to deal with the initialization dependency problem that commonly appears in edge-based approaches. Its dynamic force field, unique bidirectionality, and constrained diffusion-based level set evolution provide great freedom in contour initialization and show significant improvements in initialization independency compared to other edge-based techniques. It can handle more sophisticated topological changes than splitting and merging. It provides new potentials for edge-based active contour methods, particularly when detecting and localizing objects with unknown location, geometry, and topology.
Example results - recovering objects with internal holes and inhomogeneous intensity
| Initial Contour | Evolving Contours | Stabilized Contour |
| Geodesic |  |
| GGVF |  |
| GeoGVF |  |
| Chan–Vese |  |
| Proposed method |  |
Example results - bidirectionality (initialization on both sides of an object boundary)
| GGVF |  |
| Proposed method |  |
Example results - initialization inside an object boundary
| Basic MAC |  |
| Proposed method (lambda=1) |  |
| Proposed method (lambda=2) |  |
Example results - recovering weak/broken edges using proposed method
| Weak/Broken Edges | Initial Shape | Evolving Shapes | Stabilized Shape |
 |  |  |  |
Example results - initialization flexibility using proposed method
| Random initialization |  |
| Cross-boundary initialization |  |
| Single inside initialization |  |
| No initialization (lambda=1) |  |
| No initialization (lambda=2) |  |
Example results - recovering noisy complex shape using proposed method
| Complex shape with 50% Gaussian noise and its recovery result using force diffusion |  |
| Weak/broken edges after heavy edge thresholding and its recovery result without force diffusion |  |
Example results - segmentation of a bone CT image
| Geodesic |  |
| GGVF |  |
| GeoGVF |  |
| CPM |  |
| Chan–Vese |  |
| Proposed method (no initialization) |  |
| Proposed method (cross-boundary initialization) |  |
| Proposed method (horizontal-line initialization) |  |
Example results - segmentation of color images using proposed method
| Skin lesion image with weak edges & noise | Horse image with large color variations & weak edges | Blood vessel image with complex shape & topology, varying intensity, & weak edges | Fish swarm image with inhomogeneous color distribution, complex topology, & weak edges | |
| No initialization |  |  |  |  |
| Horizontal-line initialization |  |  |  |  |
Publications
- Xianghua Xie and Majid Mirmehdi, MAC: Magnetostatic Active Contour Model, IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI), volume 30, number 4, pages 632 - 646, IEEE CS Press, April 2008.
- Xianghua Xie, Active Contouring Based on Gradient Vector Interaction and Constrained Level Set Diffusion, IEEE Transactions on Image Processing (T-IP), volume 19, number 1, pages 154 - 164, IEEE CS Press, January 2010.
- Xianghua Xie, Si Yong Yeo, Majid Mirmehdi, Igor Sazonov, and Perumal Nithiarasu, Image Gradient Based Level Set Methods in 2D and 3D, In Deformation Models, Edited by G. Hidalgo et al., Springer, 2013.