Novel Characterization Approach for Hand Geometry-Based Verification

Abstract

The objective of this paper is to investigate a novel biometric method based on hand geometry recognition for automatically verifying the identify of a person. Hand recognition is as a promising biometric which is being used in low-level security applications for several years. Unlike other hand geometry based systems, neither the user have to undergo the inconvenience of using pages, nor the using of expensive dedicated capturing devices are employed. Hand features are extracted and used to generate some geometrical parameters that can be used to decide if two images belong to the same hand. Tests are conducted on image data sets taken from 50 users and the experimental results demonstrate that a FAR of 2% while maintaining a FRR of less than 3%, can be achieved by our method.