Palmprint recognition based on phase congruency and Two-Dimensional Principal Component Analysis

摘要

Two-Dimensional Principal Component Analysis (2DPCA) has been widely used for feature extraction in palmprint recognition. However, it is sensitive to variable illumination. In this paper, a novel method is proposed to solve this problem by combing phase congruency (PC) with 2DPCA. Our method consists of two parts. One is to extract the palmprint phase congruency features which are invariant to changes in image illumination. The other is transforming the palmprint phase congruency features into subspace by 2DPCA, which is able to classify the individual palmprint representation optimally. The design of Gabor filters for phase congruency feature extraction is also discussed. The PolyU palmprint database was used to generate the results. Experiments show that phase congruency significantly improves system performance whilst 2DPCA outperforms many existing subspace projection methods. The proposed method achieves 99.44% recognition rate on the PolyU database, and its feature extraction and matching time is 0.311s.