【摘要】 提出一种基于Gabor小波与Memetic算法的人脸识别方法MA-Gabor(Memetic Algorithm-Gabor).算法使用一组特定的Gabor小波滤波器对人脸图像重要区域进行针对性的特征提取运算,可在较短处理时间内获得更具区分能力的识别数据.为提升识别性能,MA-Gabor引入Memetic算法用于Gabor小波滤波器组的优化设计.实验结果表明,Memetic算法可获得比传统优化方法更佳的设计效果.通过将优化设计的Gabor小波滤波器组用于人脸图像的特征提取,MA-Gabor算法可取得比现有人脸识别方法更高的识别率.更多还原

【Abstract】 A Gabor wavelet and Memetic algorithm based face recognition technique named MA-Gabor is proposed in this paper.MA-Gabor can attain more distinctive features with less computational time by using particular Gabor wavelet filters in the feature extraction on specific areas of the face image.The Memetic algorithm is introduced in optimizing the Gabor wavelet filters’ design in MA-Gabor.Experimental results show that the Memetic algorithm can achieve better design than conventional optimization approaches.The MA-Gabor algorithm,using optimized Gabor wavelet filters in feature extraction,attains higher recognition rate than state-of-the-art face recognition techniques.更多还原