Hand Geometry Recognition based on optimized K-Means Clustering and Segmentation Algorithm

Abstract

Biometrics plays an important role in electronic authentication based on biological features of the human beings like hand geometry, palm print, finger print, retina and face geometry. The hardware requirements of the different electronic systems that are subset of biometric systems of categories mentioned above vary significantly. The power consumption increases with the hardware complexity used. Large power consumption leads to less portability which is not desired. Most biometric recognition systems require complex optical systems for properly capturing data without noise. The digital pre-processing of input image is an added advantage. Hand Geometry Recognition is the simplest and robust recognition system among all other recognition methods available. Hand geometry recognition involves extracting the hand which is the foreground from the background using segmentation process. Several segmentation algorithms are available. K-Means algorithm is one of the algorithms. In this work the K-Means algorithm is optimized to be used in the recognition process.