基于复杂网络和图像轮廓的形状识别方法  

A Shape Recognition Method Based on Complex Network and Image Contour

作　　者：汤晓[1] 王银河[1] 

机构地区：[1]广东工业大学自动化学院,广东广州510006

出　　处：《山西师范大学学报（自然科学版）》2013年第3期33-40,共8页Journal of Shanxi Normal University(Natural Science Edition)

基　　金：国家自然科学基金项目(61273219);广东省自然科学基金(8151009001000061);广东省自然科学联合研究基金项目(8351009001000002)

摘　　要：基于轮廓的图像识别方法具有过程简单、识别效率高等特点.但随着人脸表情变化、光照强度以及遮挡等因素的改变,提取图像轮廓形状的难度增大,从而使方法的有效性受到影响.本文提出一种基于复杂网络和图像轮廓的形状识别方法,通过提取形状图像的轮廓点,建立相应的复杂网络模型,计算相关参数来识别图像.实验表明,该方法具有对轮廓图精确度依赖性低、复杂网络规模小、阈值参数少、能有效适应边界形状改变等优点.The method based on image contour and shape can achieve high identify efficiency and simple process. But the change of the facial expression, light intensity or position of shelters will increase the difficulty in extracting the contour and reduce the recognition effect. To solve these problems, a shape recognition method based on complex network and image contour is discussed in this paper. The main idea of the approach is to ex- tract the shape image contour, model them into graphs and use complex network methodology to extract a feature vector for face recognition. Experiments show that the proposed method could reduce the reliance on the accuracy of contour extracting and effectively control the scale of the complex network with efficient power of face recogni- tion.

关 键 词：复杂网络 图像轮廓 形状识别 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]