Two-stage image segmentation based on edge and region information  

作　　者：冉鑫 戚飞虎 

机构地区：[1]Merchant Marine College, Shanghai Maritime University [2]Dept. of Computer Science and Engineering, Shanghai Jiaotong University

出　　处：《Journal of Harbin Institute of Technology(New Series)》2007年第4期533-540,共8页哈尔滨工业大学学报（英文版）

基　　金：Sponsored by Shanghai Leading Academic Discipline Project(Grant No T0603);the National Natural Science Foundation of China (Grant No60271033)

摘　　要：A two-stage method for image segmentation based on edge and region information is proposed. Different deformation schemes are used at two stages for segmenting the object correctly in image plane. At the first stage, the contour of the model is divided into several segments hierarchically that deform respectively using affine transformation. After the contour is deformed to the approximate boundary of object, a fine match mechanism using statistical information of local region to redefine the external energy of the model is used to make the contour fit the object's boundary exactly. The algorithm is effective, as the hierarchical segmental deformation makes use of the globe and local information of the image, the affine transformation keeps the consistency of the model, and the reformative approaches of computing the internal energy and external energy are proposed to reduce the algorithm complexity. The adaptive method of defining the search area at the second stage makes the model converge quickly. The experimental results indicate that the proposed model is effective and robust to local minima and able to search for concave objects.A two-stage method for image segmentation based on edge and region information is proposed. Different deformation schemes are used at two stages for segmenting the object correctly in image plane. At the first stage, the contour of the model is divided into several segments hierarchically that deform respectively using affine transformation. After the contour is deformed to the approximate boundary of object, a fine match mechanism using statistical information of local region to redefine the external energy of the model is used to make the contour fit the object＇s boundary exactly. The algorithm is effective, as the hierarchical segmental deformation makes use of the globe and local information of the image, the affine transformation keeps the consistency of the model, and the reformative approaches of computing the internal energy and external energy are proposed to reduce the algorithm complexity. The adaptive method of defining the search area at the second stage makes the model converge quickly. The experimental results indicate that the proposed model is effective and robust to local minima and able to search for concave objects.

关 键 词：active contour model region growing image segmentation affine transformation 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]