Structural Feature Measurement Using Fast VO Model for Blurred Palmprint Recognition

摘要

It is inevitable for non-contact palmprint recognition to obtain the low resolution image when capturing the image, which leads to poor recognition accuracy. In order to solve this problem effectively, a blurred palmprint recognition method based on Structure Feature (SF) is proposed in the paper. Firstly, fast Vese-Osher (VO) decomposition model is utilized to decompose blurred images in order to obtain stable feature of blurred images which is regarded as SF. Next, a non-overlapping sampling method based on Structure Ratio (SR) for SF is used to further improve recognition accuracy. Finally, Structural Similarity Index Measurement (SSIM) is used to measure the similarity of palmprints and judge the palmprint category for classification. The recognition results of proposed method are stable in the PolyU palmprint database and the Blurred-PolyU palmprint database, moreover, the Equal Error Rate (EER: 0.9069%) of proposed method is lower than other classical algorithms in Blurred-PolyU palmprint database.