A Level Set Image Segmentation Method Using Prior Region Information

• Published in: IET Conference Proceedings pp. 96 - 99
• Main Authors: Yuan-ming, Dai, Wei, Wei, Yi-ning, Lin
• Format: Conference Proceeding
• Language: English
• Published: Stevenage, UK IET 2012; The Institution of Engineering & Technology
• Subjects: Computer vision and image processing techniques; Information retrieval techniques; Interpolation and function approximation (numerical analysis); Optical, image and video signal processing; image analysis applications; image segmentation; target tracking; level set image segmentation method; image retrieval; statistic distance maximization criterion; segment accuracy improvement; curve fitting; curve evolution energy function; prior region information; Euler-Lagrange equation; dimidiate image
• ISBN: 9781849195379; 1849195374
• DOI: 10.1049/cp.2012.0929

Image Segmentation is a promising technology and very important for image retrieval, target tracking and image analysis applications. The existing region-based level set image segmentation methods which obey a criterion of maximizing the statistic distance of sub-region are mostly on the assumption of dimidiate image, often leading to unsatisfied segmentation results. In order to overcome this deficiency, this paper presents a novel level set image segmentation method using prior region information to improve the segment accuracy. The paper firstly introduces how the prior information is adopted into curve evolution energy function. Then the associated Euler-Lagrange equation which is the most important step is deduced. Finally, the implementation details for the proposed novel method are presented. Extensive experimental results show that the novel level set image segmentation method using prior region information presented in this paper has achieved higher accuracy compared with existing methods.