文章基本信息

标题：Image Segmentation Using Grabcut Based on Anisotropic Diffusion
本地全文：下载
作者：Anju Rachel Oommen ; Nathelda Mary Navina
期刊名称：International Journal on Computer Science and Engineering
印刷版ISSN：2229-5631
电子版ISSN：0975-3397
出版年度：2011
卷号：3
期号：03
页码：1307-1311
出版社：Engg Journals Publications
摘要：Image segmentation is a technique used to extract the foreground object from background image. In this paper Grabcut framework is used to extract the object from the background. Color and Texture feature is used as discriminating feature for object extraction. Region based image segmentation is done in this paper. For Texture feature extraction, Structure Tensor is used for finding the correct boundary/edge of the needed object. Smoothing is done using Anisotropic diffusion which will preserve the edges from dislocation. If dislocation in edges arises, it is difficult to get proper segmentation. In this paper, details of anisotropic diffusion is given and it tells how the image quality will be increased using this method.
关键词：Grabcut; Structure Tensor; Diffusion;Graphcut;MSSIM.