This paper investigates novel LBP-guided active contour approaches to texture segmentation. The Local Binary Pattern (LBP) operator is well suited for texture representation, combining efficiency and effectiveness for a variety of applications. In this light,
IntroductionTexture segmentation methods based on active contour approaches have received considerable attention over the past few years (Theodoridis and Koutroumbas, 2006;Paragios and Deriche, 1999;Lehmann et al, 2001;Aujol et al, 2003;Rousson et al, 2003;Sagiv et al, 2004;Huang et al, 2004;He et al, 2004;Allili et al, 2004;Pujol and Radeva, 2004;Lee et al, 2005), by exploiting advances in the active contour research such as contour smoothness, noise robustness and topological adaptability. This emerging trend in the area of texture segmentation has been reinforced by the vector formulation of recent active contour approaches Sandberg and Chan, 2005) introduced to provide a natural platform for the embedment of textural features. Such methods constitute an essential first step in computer vision applications, which are as diverse as medical image analysis, industrial monitoring of product quality, content-based image retrieval and remote sensing.The main notion of the active contour approach to texture segmentation relies on the deformation of initial contours towards the boundaries of image regions to be segmented. The deformation is realized by minimizing an energy functional, designed so that its local minimum is reached at target boundaries. Active contour models lead to