Font Size: a A A

Application Research Of Active Contour Model In Image Segmentation

Posted on:2009-03-28Degree:MasterType:Thesis
Country:ChinaCandidate:B YaoFull Text:PDF
GTID:2178360272978143Subject:Computer system architecture
Abstract/Summary:PDF Full Text Request
Image segmentation is one of the elementary problems in image understanding and active vision. As an algorithm guided by high-level knowledge, active contour model is a hotspot of research in the field of image understanding.In this paper, we first introduce the common methods of image segmentation, then we discuss parametric active contour model and geometric active contour model, finally, in the light of the shortcoming of GVF Snake model and GGVF Snake model that can not approximate the real contour of a wide deep concave object and an object like a bottle, we propose a subsection contour detection algorithm.New algorithm get the whole contour of a wide deep concave object and an object like a bottle partly according to the critical points outside the object and the stagnation point region in the GGVF field of the object. First, get parts of contour of the object using GGVF Snake model. And then, set initial contour which includes critical points outside the object in the GGVF field of the object automatically according to false contour in the stagnation point region and get the rest contour of the object. Finally, the two parts of contour are fused into one contour.
Keywords/Search Tags:Image segmentation, Active contour model, Gradient vector flow, Critical point, Object like a bottle
PDF Full Text Request
Related items