| 基于独立分量分析和神经网络的人脸识别方法 |
| |
| 引用本文： | 王明祥,莫玉龙. 基于独立分量分析和神经网络的人脸识别方法[J]. 计算机工程与应用, 2004, 40(5): 4-6,25 |
| |
| 作者姓名： | 王明祥 莫玉龙 |
| |
| 作者单位： | 上海大学通信与信息工程学院,上海,200072;上海大学通信与信息工程学院,上海,200072 |
| |
| 基金项目： | 国家自然科学基金资助项目(编号:69875009) |
| |
| 摘 要： | 提出了一种独立分量分析(ICA)和BP神经网络相结合的人脸识别方法(ICABP法),其中人脸图象独立基的提取采用快速独立分量分析方法(FastICA),BP网络采用改进的三层网络结构。该方法将ICA的空间局部特征提取功能和BP神经网络的自适应功能有机地结合起来,增强了系统的鲁棒性。实验表明,ICABP法明显提高了人脸的识别率。
|
| 关 键 词： | 独立分量分析 主分量分析 神经网络 人脸识别 |
| 文章编号： | 1002-8331-(2004)05-0004-03 |
Face Recognition Method Based on Independent Component Analysis and Neural Network |
| |
| Wang Mingxiang Mo Yulong. Face Recognition Method Based on Independent Component Analysis and Neural Network[J]. Computer Engineering and Applications, 2004, 40(5): 4-6,25 |
| |
| Authors: | Wang Mingxiang Mo Yulong |
| |
| Abstract: | A new face recognitio n method combining independent component analysis(ICA)and BP neural network is proposed in this paper.The FastICA algorithm is used to derive the independent sources out of the face images and the improved BP neural network with three layers is used to classify faces.In this paper,the local features extraction o f ICA and the adaptability of BP neural network are combined reasonably.The rob ustness of the system is enhanced greatly.The experiment results show that the face recognition rate of ICABP method is improved obviously. |
| |
| Keywords: | Indepen dent component analysis principle component analysis neural network face r ecognition |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
