一种基于最小风险检测ART网分类器的设计与人脸识别

A Classifier Design Based on ART Network of Input Venture Detection and Face Recognition

本文通过对传统的自适应共振网络(ART)的研究,分析了其优缺点及应用领域,在此基础上提出了一种适合训练样本数较少、样本特征值维数较高的图像目标分类的最小风险检测 ART 网络,并重点讨论了该网络在结构上的改进和运行原理,以及基于该网络的分类器的设计及其算法实现过程。通过对 ORL 人脸库和 Yale 人脸库的图像样本仿真结果表明,应用该分类方法进行人脸图像分类识别,同时获得了较高的分类速度和分类效果。

Based on the study on the transient Adaptive Resonance Theory（ART）and their merits and defects, an im proved ART network of input venture detection adapted to classification of less training stylebooks and high dimension images is presented in this paper. This paper firstly discusses the improvement of the structure of the network and its working principle in detail. Then the classifier based on this network and the process of the algorithm＇s realization is emphatically discussed. Through the experiment on the libraries of ORL and Yale human face, satisfactory results and high classification rate have been obtained by analyzing the course of the algorithm with the designed image classifier, especially in the field of face image classification.