A hand geometry verification system relating finger length, position of the finger base and palm center

Abstract

Fingerprints and Iris scans have been the most popular form of biometrics for a long time. A more niche form of biometrics – based on geometric features has existed since the 1980’s. The ideas were promising due to less intrusiveness and low storage requirements but had hardly seen use due to lower accuracy. In recent years however, geometric techniques are attracting attention as a result of better hardware and improved algorithms. In this paper, an algorithm to verify a person’s identity based on the hand shape is presented. The paper gives an introduction to hand geometry then surveys existing work in the field. The main algorithm, which calculates the features to authorize and identify a person by deriving relations from attributes like finger length, coordinates of the finger base and palm center is explained in detail in the methodology section. The system is then tested on a locally collected dataset with 30 images from 7 different individuals. The results of the testing are provided in the experiment section, followed by a conclusion.