Hand Geometry Based Person Verification System

Abstract

Hand geometry which is the famous biometrics system is utilized in the various identification systems with different feature estimation methods. Hand geometry plays a vital rule in the biometric authentication and security application. For this regard, this paper presents a biometric system for person verification based on the hand geometric features. The hand geometric features are extracted from the upper palm including four fingers of the right hand. The main aim of this proposed system is to reduce the feature and database size and improve the performance of the system. For that, initially, top four fingertip points and two corner valley points from the right hand four fingers are estimated. After that, eight distance edges from these fingertip and valley points are computed. From these eight distance edges, three triangles are estimated. These three triangle areas are used as a three feature and stored in the database as the feature vector. Finally, the test candidate hand feature is compared with the predefine database feature vector through the Euclidian distance metric. The proposed verification system is tested with the own dataset size of 250 user's images i.e., 1250 hand images and results show the momentous improvement with compare to the existing state of the art.