Robust Person Identification through Palmprint Verification
Abstract
Palm print has become one of the most important parts of biometric system for person verification since it has numerous features like principle lines, wrinkles, ridges, minutiae points, singular points, and texture pattern for representation. The Scale Invariant Feature Transform operator of human hand is used for this purpose. Usually, low cost scanners are used for the purpose of scanning. Usually, the hand which is being kept over a scanner is subjected to translation and rotation. Various techniques are being used to match the palm prints. However, there are a lot of problems in the matching of palm prints. Features are extracted using Scale Invariant Feature Transform (SIFT). This work proposes a Palm print verification technique which sets a threshold value based on the OSTU Algorithm and uses Mahalanobis distance for the verification of Palm prints. The OSTU method is one of the applied methods of image segmentation in selecting threshold automatically for its simple calculation and good adaptation. Further on the usage of the Mahalanobis distance enhances the verification process.
Keywords Binarization, OSTU Algorithm Scale Invariant Feature Transform (SIFT) , Mahalanobis Distance