基于神经网络的二维主成分分析人脸识别算法_邵天双.pdfVIP

第 28卷第 3期 计算机应用与软件 Vol128No. 3 2011年 3月 Computer Applications and Sofwt are M ar. 2011 1 2 邵天双 金丽华 1 ( 518055) 2 ( 15008 ) 提出基于神经网络的二维主成分分析人脸识别算法通过图像的 处理改善了图像的质量, 提高了图像的亮度和对比 度, 降低了图像的维度, 然后利用二维主成分分析方法进行人脸关键特征的提取, 并将该特征作为神经网络的输入, 用改进的神经网 络作为分类器, 并通过实验证明了算法的有效性和可行性 人脸识别 神经网络 二维主成分分析 小波变换 2DPCA FACE RECOGNITION ALGORITHM BASED ON NEURAL NETWORK 1 2 Shao T ianshuang Jin L ihua 1 (H arbin Institute of Technology Shenzhen Graduate School, Shenzhen518055, Guangdong, China) (N o. L and Surveying Team, State Bureau of Surverying andM app ing, Harbin 150086, H eilongj iang, Ch ina) Abstract Neural newt ork-based 2DPCA hum an face recognition algorithm ispresented. By pre-processing the mi age, the quality of mi age is enhanced, and the contrast grade and brightness of gray mi age are mi proved too, the d mi ension of the mi age is reduced asw ell. Then the two-d mi ensional principal com ponent analysis is applied to extract key featu res of hum an face, the features are em ployed as the input of neural network, and the mi proved neural network is used as classifier. Expermi ental resu lts show that the new algorithm is effective and feasib le. K eywor s Face recogn ition Neural network Tw o-dmi ensional principal componen t analys is (2DPCA) W avelet transform , 0 1. 1 , , , , , , ,