Designing Algorithms for Human Vein features Recognition System

Abstract

Biometrics plays an important role in identifying the humans based on their various newlineunique biological characteristics like fingerprint, palm print, faces, hand geometries, newlineDNA, retina, iris, voice, gait and signatures. But, these biometric characteristics newline(traits) are not robust against the forgery and can easily be misused to fool the newlinebiometric system. Recently, vein pattern based biometric has emerged as promising newlinealternative and attracted an attention of many researchers. This vein biometric uses newlinehuman s vein (vascular) pattern for the personal identification based on the fact that newlineevery individual has a unique veins pattern in their body. Since, vein pattern is an newlineinherent trait, it is robust against the forgery and does not affect due to external factors newlinesuch as dirt, moisture and injuries. Therefore, vein pattern has become a most newlinepromising and efficient biometric recognition due to its accuracy, security and newlineconvenience. The vein pattern image can t be captured by using normal cameras, newlinetherefore near infrared cameras are used for capturing the veins pattern image from newlinethe human finger, palm and dorsal. As infrared light rays passes through the skin, newlinehemoglobin in the blood absorbs rays and reflects the vein lines as dark lines and newlineother part as bright. The vein pattern image capturing process from human finger is newlineefficient and convenient due to its small device size and cost, hence finger veins newlinepattern is mostly preferred. Several authors have proposed different approaches for newlinefinger vein features extraction and their recognition but still there is a need of newlinedeveloping a novel approach that can extract more accurate and discriminate features newlinefor finger vein recognition with less error rate. This research work presents an newlineapproach for the personal identification using local and global features of human newlinefinger vein pattern images. In order to carry out this research work, finger vein pattern newlinelocal features are extracted using Local Directional Code method and the global newlineiv newlinefeatures extracted using Discrete W