Robust level set method for computer vision

Jia-rui Si; Xiao-pei Li; Hong-wei Zhang

doi:10.1117/12.664288

20 February 2006 Robust level set method for computer vision

Jia-rui Si, Xiao-pei Li, Hong-wei Zhang

Proceedings Volume 6041, ICMIT 2005: Information Systems and Signal Processing; 60410B (2006) https://doi.org/10.1117/12.664288
Event: ICMIT 2005: Merchatronics, MEMS, and Smart Materials, 2005, Chongqing, China

Abstract

Level set method provides powerful numerical techniques for analyzing and solving interface evolution problems based on partial differential equations. It is particularly appropriate for image segmentation and other computer vision tasks. However, there exists noise in every image and the noise is the main obstacle to image segmentation. In level set method, the propagation fronts are apt to leak through the gaps at locations of missing or fuzzy boundaries that are caused by noise. The robust level set method proposed in this paper is based on the adaptive Gaussian filter. The fast marching method provides a fast implementation for level set method and the adaptive Gaussian filter can adapt itself to the local characteristics of an image by adjusting its variance. Thus, the different parts of an image can be smoothed in different way according to the degree of noisiness and the type of edges. Experiments results demonstrate that the adaptive Gaussian filter can greatly reduce the noise without distorting the image and made the level set methods more robust and accurate.

Citation Download Citation

Jia-rui Si, Xiao-pei Li, and Hong-wei Zhang "Robust level set method for computer vision", Proc. SPIE 6041, ICMIT 2005: Information Systems and Signal Processing, 60410B (20 February 2006); https://doi.org/10.1117/12.664288

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available