Abstract:

Palm Vein Recognition is one among the biometric authentication methods. It works by recognizing the unique patterns of veins in the palms of the people. Generally, veins carry deoxygenated blood from tissues to the heart. In palm vein recognition, the near-infrared light is illuminated into the palm emitted by the palm scanner. This near-infrared light is absorbed by the deoxygenated blood flowing through the veins which reduces its ability to reflect back the light. This causes the veins to appear as black patterns and is captured as an image by the scanner. These images are processed using various algorithms and compared with the data in the database and authenticates the individuals. In this work, we are using a Mutual Foreground based Local Binary Pattern algorithm for the processing of palm vein images. Local Binary Pattern (LBP) is an efficient texture representation of an image, but when used to describe the sparse texture in palm vein images, the discrimination ability is diluted, leading to lower performance, especially for contactless palm vein matching. In this work, an improved Mutual Foreground LBP method is used for achieving a better matching performance for contactless palm vein recognition. First, the palm print images are pre-processed. The aim of pre-processing is to enhance the features of the palm image for better analysis and further processing. The k-means algorithm is utilized for removing noise, texture extraction, improves accuracy and robustness. Later we will find the Region of Interest (ROI) which plays a vital role in maintaining tolerance. Then, a Mutual Foreground based LBP matching strategy is adopted for identifying the similarities on the basis of extracted palm veins. LBP contains relative values obtained by comparing each pixel with its neighboring pixels. In addition, we adopted the matched pixel ratio (MPR) to find the best matching region (BMR), which can further improve the performance of LBP matching. Third, the matching score obtained in the process of finding the BMR was fused with the results of LBP matching to further improve the identification performance. Then, an improved Chi-square distance is proposed to increase the computational efficiency. Finally, the person is authenticated and shows whether he/she is allowed to access the device.

Description:

Palm Vein Recognition, Local Binary Pattern, mutual foreground, infrared light, palm, palm veins, patterns,pre-processing