A Novel Approach for Pattern Recognization Using Neural Network of Hand Biometrics

Hand geometry is considered to achieve medium security with lesser cost along with several other advantages as compared to other available techniques. The physical dimensions of a human hand known as hand or palm geometry, contains information that is capable of authenticating the identity of an individual. The hand geometry based identity verification system is being widely used in various applications like access control, time and attendance, point-of-scale, anti-pass back and interactive kiosks etc. The goal of a biometric verification system consists in deciding whether two characteristics belong to the same person or not. The present work presents a biometric user recognition system based on hand geometry. It explores the features of a human hand, extracted from a color photograph which is taken when the user is asked to place his/her hand on a platform especially designed for this task.

Different pattern reorganization techniques have been used for classification and verification. It consists of a database where all the information about the authenticated users is stored. The system extracts the features from a test image and compares it with the stored information on the database. The experimental results show that the proposed system has an encouraging performance. The false acceptance rate are reduced down to 0.02, respectively. Experimental result, up to 95 percent rate of success in classification, will show the possibility of using the system in medium/high security environment with less cost.