Vol.27 No.4 JOURNAL OF ELECTRONICS (CHINA) July 2010 REGION-BASED ACTIVE CONTOUR DRIVEN BY GLOBAL INTENSITY FITTING ENERGY 1 Tian Yun Zhou Mingquan Wu Zhongke Wang Xingce (College of Information Science & Technology, Beijing Normal University, Beijing 100875, China) Abstract In this paper, we present a novel region-based active contour model based on global in- tensity fitting energy in a variational level set framework. Meanwhile, an internal energy term is in- troduced, and it forces the level set function to be close to a signed distance function. Image global information utilized efficiently makes the proposed model insensitive to noise, and the introduced penalty term can avoid the costly re-initialization for the evolving level set function, which not only speeds up the contour evolvement, but also improves accuracy of the final contour. Comparisons with other classical region-based models, such as Chan-Vese model and Region-Scalable Fitting (RSF) model, show the advantages of our model in terms of efficiency and accuracy. Moreover, the model is robust to noise. Key words Region-based active contours; Internal energy; Global energy; Level set; Re-initialization CLC index TP391.4 DOI 10.1007/s11767-011-0344-z I. Introduction In recent years, active contour models have been one of the most successful methods for boundary extraction and image segmentation. The existing active contour models can be categorized into two classes: edge-based models [1–5] and re- gion-based models [6–8] . Edge-based models utilize image gradient to stop the evolving contours on the object boundaries. In practice, it is difficult to control the motion of the contour. Region-based models aim to identify each region of interest by using a certain region descriptor to guide the mo- tion of the active contour, and do not utilize the image gradient and thus have better performance for the image with weak object boundaries. In addition, region-based models are significantly less sensitive to the location of initial contours. One of the most popular region-based models are Chan- Vese model [6] and its extended ones [7–10] . These methods rely fully on the global information of the 1 Manuscript received date: December 15, 2009; revised date: April 29, 2010. Supported by the State Key Program of National Natural Science of China (No. 61003134, 60736008), the National Natural Science Foundation of China (No. 60803082) and the Key Program of Natural Science of Beijing (No. 4081002). Communication author: Tian Yun, born in 1980, male, Ph.D.. Beijing Normal University, Beijing 100875, China. Email: tyun0619@gmail.com. image, instead of its local gradient, and they can realize global optimization of the image segmenta- tion through minimization Mumford-Shad energy functional. Unfortunately, the methods based on Chan-Vese model are computationally inefficient. Moreover, they may not be able to re-initialize the level set function to a signed distance function when the level set function is far away from a signed distance function. In practice, the evolving level set function can deviate greatly from its value as signed distance in a small number of iteration steps, especially when the time step is not chosen small enough. Li, et al. [7,8] proposes an algorithm based on minimization of Region-Scalable Fitting (RSF) energy for image segmentation with inten- sity inhomogeneities. We can call this model RSF for short. RSF relies heavily on extracted intensity information in local regions to guide the motion of the contour, which thereby enables the model sen- sitive to the noise. In this paper, we propose a different region- based active contour model which is based on the global information, instead of the gradient and local information of the image, for the stopping process. The basic idea is to introduce a global energy term and an internal energy term in a variational formulation. The variational energy functional model is composed of the global energy, the signed distance function, the length and the