Analysis of Palm Vein Features through An Efficient Algorithm

Abstract

The technique of palm vein image feature extraction addresses the issues relating to newlinethe optimality of decision making in the person identification framework. The newlineadvances in feature extraction algorithm enables to perform person identification by a newlinemore reliable and accurate way. The thesis discusses feature extraction techniques in newlinedetail and describes the theoretical and experimental work done to show its reliability newlineand validity. newlineThe current work intends to analyze contactless palm vein image features by different newlinealgorithms and look into its discriminative energy as a biometric application. Since newlinepalm vein image features involve multi-dimensional data set, the algorithm built must newlinebe able to optimize the palm vein image features. The present research work is newlineintended to elaborate the following consideration namely investigating appropriate newlinepreprocessing techniques for palm vein images, investigating image quality evaluation newlinealgorithms for palm vein images and identifying different combination of palm vein newlineimage feature optimization and the extraction algorithms. The research work extends newlineits investigation in identifying effective image classification and recognition method. newlineAlso, testing the palm vein image features extracted on hardware to check the newlinefeasibility of the algorithms helps in designing possible real time system development newlinefor secure authentication. newlineThe primary objective of the present work is to explore how to improve the feature newlineextraction process of palm vein image in order to construct reliable palm vein newlinerecognition systems. The aim of the proposed thesis work is to develop a theoretical newlineand practical basis for enhancing the performance of palm vein pattern recognition newlinesystem using advanced methods and machine learning algorithms. The present work newlinestarts with the different preprocessing techniques that are carried out to overcome any newlinegeometrical variations such as image orientation, shearing, frequency estimation and newlinegray level transformation error. Some of the objecti