A HYBRID FACE RECOGNITION APPROACH USING LOCAL FUSION OF COMPLEX DUAL-TREE WAVELET COEFFICIENTS AND RIDGELET TRANSFORM

In this paper, we propose novel face recognition method based on local appearance feature extraction using hybrid mode of local ridge- let and fused dual-tree complex wavelet transform (DT-CWT). It provides a local multiscale description of images with good directional selectivity, effective edge representation and invariance to shifts and in-plane rotations. In the dual-tree implementation, two parallel dis- crete wavelet transform (DWT) with different lowpass and highpass filters in different scales are used. The linear combination of sub- bands generated by two parallel DWT is used to generate 6 different directional subbands with complex coefficients. It is insensitive to illumination variations and facial expression changes. 2-D dual-tree complex wavelet transform is less redundant and computationally efficient. The fusion of local DT-CWT coefficients of detail subbands and local Finite Ridgelet Transform (FRIT) coefficients of approxi- mate subbands of DT-CWT are used to extract the facial features which improve the face recognition with small sample size in less computation. The local features based methods have been success- fully applied to face recognition and achieved state-of-the-art per- formance. Normally most of the local appearance based methods the facial features are extracted from several local regions and concate- nated into an enhanced feature vector as a face descriptor. In this approach we divide the face into several (m×m) non-overlapped paral- lelogram blocks instead of square or rectangle blocks. The local mean and standard deviation of hybrid FRIT and fused DT-CWT coeffi- cients are used to describe the face image. Experiments, on two well- known databases, namely, Yale and ORL databases, shows the Local hybrid FRIT and fused DT-CWT approach performs well on illumi- nation, expression and perspective variant faces with single sample compared to PCA and global DT-CWT. Furthermore, in addition to the consistent and promising classification performances, our pro- posed Hybrid Local FRIT and fused DT-CWT based method has a really low computational complexity.