自动面积梯度分割算法及其在耳形图像中的应用  

Auto-area-grads Segmentation Algorithm and Its Application on Ear Image

作　　者：黎燕[1] 樊小平[1] 李刚[1] 

机构地区：[1]中南大学信息科学与工程学院,长沙410075

出　　处：《计算机工程与应用》2006年第6期55-57,共3页Computer Engineering and Applications

摘　　要：提出一种实现抗噪保形的自动面积梯度分割算法。这种算法首先通过将图像分层并且利用最大熵原理确定面积阈值,然后在每一层进行面积开或闭运算,同时运用梯度算子,最后合并所有层。将这种分割算法应用到耳形图像的分割,外耳形状的特征点和特征线明显的分割出来,非常利于识别;在其它类型图像中也可取得满意效果。A method of area-grads segment for anti-noise and keeping contours is presented.The algorithm is that the image is divided into many level sets and the theory of maximal entropy is used to gain the area threshold,then in level sets area-opening or area-closing is applied,meanwhile grads operators is used.At last all level sets are added to form a new image.This segment algorithm is applied in ear image and feature points and curves of outside ear are ext racted,which helps to recognize.The algorithm also gains good result in other type image.

关 键 词：面积形态学 面积阈值 梯度算子 图像分割 耳形图像 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]