14 March 2013 Rectification-adapted snake for complex-boundary segmentation in noisy images
Author Affiliations +
Proceedings Volume 8768, International Conference on Graphic and Image Processing (ICGIP 2012); 87687H (2013) https://doi.org/10.1117/12.2010100
Event: 2012 International Conference on Graphic and Image Processing, 2012, Singapore, Singapore
Abstract
In this paper, a contour-fitness improved adaptive snake, namely, edge-conducted rectification-adapted snake (ECRA-snake) is proposed for segmenting complex-boundary objects in the noisy image. The ECRA-snake includes a main ingredient called edge-conducted evolution (ECE), where the adaptations of model coefficients can accommodate ECE itself to the characteristics of salient edges for better curve fitting in tracking. Following ECE, a direction-induced rectification evolution (DIRE) will correct boundary-unmatched snake fragments by handling the initial direction and the tensile-force weighting of unqualified snaxels in this snake re-evolution. Simulation results demonstrate that the proposed ECRA-snake can obtain better object-boundary coincidence than the Gradient Vector Flow (GVF) model in segmenting the complex-boundary object from noisy images.
© (2013) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Din-Yuen Chan, Roy Chaoming Hsu, Pang-Hao Wu, Cheng-Ting Liu, "Rectification-adapted snake for complex-boundary segmentation in noisy images", Proc. SPIE 8768, International Conference on Graphic and Image Processing (ICGIP 2012), 87687H (14 March 2013); doi: 10.1117/12.2010100; https://doi.org/10.1117/12.2010100
PROCEEDINGS
6 PAGES


SHARE
RELATED CONTENT

BTF Potts compound texture model
Proceedings of SPIE (March 13 2015)
Free-form quadric surface patches
Proceedings of SPIE (August 01 1990)
Adaptive design of visual perception experiments
Proceedings of SPIE (April 22 2010)
Fast image completion
Proceedings of SPIE (July 31 2006)

Back to Top