基于变换域特征提取和模拟退火法特征选择的人脸识别
|
摘要
为了提高人脸识别系统的性能,提出DWT双边子带频域特征提取和模拟退火优化算法的特征选择。首先,运用DWT、DFT和DCT组合变换,用于人脸图像表情、姿态、平移和光照不变特征的有效特征提取。人脸图像DWT变换后,选择近似系数分量和水平系数分量[CA CH]为小波特征,经DFT变换后,利用四椭圆模板取出DFT低频高幅值系数,经DCT压缩得到人脸图像的特征系数。其次,利用模拟退火优化算法进行特征选择,在特征系数空间搜索特征子集进行人脸识别。实验仿真说明了该方法的有效性。
Two core techniques are proposed in the paper to improve the performance of face recognition system: DWT dual_sub band frequency domain for feature extraction and simulated annealing optimization for feature selection. First,it is combined DWT、DFT and DCT methods to extract facial expressions,translation and illumination invariant features. The wavelet features [CA CH]are selected using approximation and horizontal coefficients of the DWT of a face images. Low frequency high amplitude components are achieved with quadruple ellipse mask after DFT. DCT feature coefficients are obtained with DCT compression in the end. Then,The Simulated Annealing Algorithm is used for features selection and the feature subset is searched for the recognition in the space. The experiment simulation shows the validity of the method.
Two core techniques are proposed in the paper to improve the performance of face recognition system: DWT dual_sub band frequency domain for feature extraction and simulated annealing optimization for feature selection. First,it is combined DWT、DFT and DCT methods to extract facial expressions,translation and illumination invariant features. The wavelet features [CA CH]are selected using approximation and horizontal coefficients of the DWT of a face images. Low frequency high amplitude components are achieved with quadruple ellipse mask after DFT. DCT feature coefficients are obtained with DCT compression in the end. Then,The Simulated Annealing Algorithm is used for features selection and the feature subset is searched for the recognition in the space. The experiment simulation shows the validity of the method.
引文
[1]Sun T H,Tien F C.Using Bp neural network for face recognition with 2D+3D hybrid information[J].Expert Systems with Applications,2008,35(1-2):361-372.
[2]Yang J,Ling X F,Zhu Y T,et al.Face detection and recognition system in color image series[J].Mathematics and computer in simulation,2008,77(5-6):531-539.
[3]Thiyagarajar R,Arulselvi S,Sainarayanan G.Gabor feature based classification using statistical models for face recognition[C]//Procedia computer science,2010,2:83-93.
[4]Pong K H,Lam K M.Multi-resolution feature fusion for face recognition[J].Pattern Recognition,2014,47(2):556-567.
[5]Sengur A.Wavelet transform and adaptive neuro-fuzzy inference for color texture classification[J].Expert Systems with Applications,2008,34(3):2120-2128.
[6]Gumus E,Kilic N,Sertbas A,et al.Ucan.Evaluation of face recognition techniques using PCA、wavelets and SVM[J].Expert system with applications,2010,37(9):6404-6408.
[7]Rangaswamy Y,Raja K B,Venugopal K R.Face Recognition Using Fusion of DTCWT and FFT features[C]//Procedia Computer Science,2015,54:809-817.
[8]Balasubramanian M,Palanivel S,Ramalingam V.Real time face and mouth recognition using radial basis function neural networks[J].Expert Systems with Applications,2009,33(3):6879-6888.
[9]Ajit Krisshna N L,Kadetotad D V,Manikantan K,et al.Face recognition using transform domain feature extraction and PSO-based feature selection[J].Applied soft computing,2014,22:141-161.
[10]Gonzalez R C,Woods R E.Digital Image Processing[M].2版.北京:电子工业出版社,2007:35-40.
[11]Kumar S.Neural Networks[M].北京:清华大学出版社,2006:68-70.
[12]Haykin S.Neural Networks and Learning Machines[M].3版.北京:机械工业出版社,2011:22-30.
[13]Theodoridis S,Koutroumbas K.Pattern Recognition[M].4版.北京:机械工业出版社,2009:34-36.
[2]Yang J,Ling X F,Zhu Y T,et al.Face detection and recognition system in color image series[J].Mathematics and computer in simulation,2008,77(5-6):531-539.
[3]Thiyagarajar R,Arulselvi S,Sainarayanan G.Gabor feature based classification using statistical models for face recognition[C]//Procedia computer science,2010,2:83-93.
[4]Pong K H,Lam K M.Multi-resolution feature fusion for face recognition[J].Pattern Recognition,2014,47(2):556-567.
[5]Sengur A.Wavelet transform and adaptive neuro-fuzzy inference for color texture classification[J].Expert Systems with Applications,2008,34(3):2120-2128.
[6]Gumus E,Kilic N,Sertbas A,et al.Ucan.Evaluation of face recognition techniques using PCA、wavelets and SVM[J].Expert system with applications,2010,37(9):6404-6408.
[7]Rangaswamy Y,Raja K B,Venugopal K R.Face Recognition Using Fusion of DTCWT and FFT features[C]//Procedia Computer Science,2015,54:809-817.
[8]Balasubramanian M,Palanivel S,Ramalingam V.Real time face and mouth recognition using radial basis function neural networks[J].Expert Systems with Applications,2009,33(3):6879-6888.
[9]Ajit Krisshna N L,Kadetotad D V,Manikantan K,et al.Face recognition using transform domain feature extraction and PSO-based feature selection[J].Applied soft computing,2014,22:141-161.
[10]Gonzalez R C,Woods R E.Digital Image Processing[M].2版.北京:电子工业出版社,2007:35-40.
[11]Kumar S.Neural Networks[M].北京:清华大学出版社,2006:68-70.
[12]Haykin S.Neural Networks and Learning Machines[M].3版.北京:机械工业出版社,2011:22-30.
[13]Theodoridis S,Koutroumbas K.Pattern Recognition[M].4版.北京:机械工业出版社,2009:34-36.