Palmprint Recognition Using Wavelet Decomposition and 2D Principal Component Analysis

In this paper, a novel method using wavelet decomposition and 2D Principal component analysis (2DPCA) for palmprint recognition is presented. Firstly, 2D wavelet transform is adopted to obtain different level of wavelet coefficients of the original palmprint image; secondly 2DPCA is applied on the low-frequency that contains most discrimination information of the original palmprint image. One criterion that not all PCs are useful for palmprint recognition is demonstrated and a rule for selecting 2D PCs is proposed. Lastly, this algorithm is tested on the PolyU palmprint image database and the experimental result is encouraging and achieves comparatively high recognition accuracy and more computationally efficient than using other feature extraction techniques such as principal component analysis and independent component analysis