一种用于人脸识别的有监督核化多类多流形ISOMAP算法.pdfVIP

第27 卷 第5 期 控 制 与 决 策 2012 年 5 月 Vol. 27 No. 5 Control and Decision May 2012 文章编号: 1001-0920 (2012) 05-0713-06 一种用于人脸识别的有监督核化多类多流形ISOMAP 算法 1 1 2 1 1 程起才 , 王洪元 , 吴小俊 , 冯 燕 , 刘锁兰 (1. 常州大学信息学院，江苏常州213164 ；2. 江南大学物联网工程学院，江苏无锡214000) 摘 要: 在核等测距映射(kernel ISOMAP) 和多类多流形ISOMAP 算法的基础上, 提出一种针对人脸识别任务的有 监督核化多类多流形ISOMAP 算法. 该算法保持了kernel ISOMAP 算法的泛化特性, 同时又能完成分类任务, 解决 了ISOMAP-C 在对具有高维小样本特性的人脸数据集识别时, 所要调整的神经网络权值数目将随输入维度呈指数增 长, 且易出现过拟合现象的问题. 在多种人脸数据集上的实验结果表明了该算法的有效性, 且对训练样本集的大小有 较好的鲁棒性. 关键词: 核等测距映射；多类多流形；人脸识别；小样本 中图分类号: TP311 文献标识码: A A supervised kernelized multi-class multi-manifold ISOMAP algorithm for face recognition 1 1 2 1 1 CHENG Qi-cai , WANG Hong-yuan , WU Xiao-jun , FENG Yan , LIU Suo-lan (1. School of Information，Changzhou University，Changzhou 213164，China ；2. School of Internet of Things Engineering ，Jiangnan University ，Wuxi 214000 ，China ．Correspondent ：WANG Hong-yuan ，E-mail: hywang@) Abstract: Based on kernel ISOMAP and multi-class multi-manifold ISOMAP, supervised kernelized multi-class multi- manifold ISOMAP is proposed specially for the facial recognition, which not only preserves the generalization property of kernel ISOMAP, but also works in a supervised manner and solves the problem that the number of neural network’s weights to be tuned increases exponentially with input dimension and easily occurs overfitting while using ISOMAP-C for face recognition. The experimental results on some face datasets show that the proposed method is effective and robust for the number of training sample. Key words: kernel I