【作者】 尹克重； 龚卫国； 李伟红； 梁毅雄；

【Author】 Yin Kezhong Gong Weiguo Li Weihong Liang Yixiong (Key Lab for Optoelectronic Technology & System of MOE, Chongqing 400044, China)

【机构】 重庆大学光电技术及系统教育部重点实验室；

【摘要】 独立成分分析(ICA)最早应用于盲信号分离,近年来也广泛地应用于模式识别领域。提出一种基于小波变换(WT)及独立成分分析(ICA)的人脸识别方法。在建议的识别方法中,首先利用小波变换将人脸图像分解成不同的频率子带,对其中包含主要信息的低频子带运用ICA求取基向量,然后基于这些基向量张成的子空间实现识别。详细讲述了该方法的实验过程,实验结果表明,该方法可以取得较好的识别结果。更多还原

【Abstract】 Independent component analysis (ICA) is an approach first used in blind source separation and now widely applied in pattern recognition domain. It introduces a new method based on wavelet transform (WT) and ICA to face recognition. In the proposed method, firstly an image is decomposed by using WT into different frequency subbands, and then ICA is applied on wavelet subbands to get the basis vector, which includes the main information of original image, lastly face recognition is implemented with the subspace comprised by these basis vectors. The experiments on ORL face database indicate the effectiveness of the proposed method.更多还原