Vein matching using line pattern

• Main Author: Li, Wei Hui
• Other Authors: Li Fang
• Format: Final Year Project
• Language: English
• Published: 2015
• Subjects: DRNTU::Engineering::Electrical and electronic engineering
• Online Access: http://hdl.handle.net/10356/62700
• Institution: Nanyang Technological University

Authentication technologies have been improving radically over the years, from the traditionally known password and smartcard authentications, to today’s biometric authentication methods. Biometric authentication is an automated recognition of individual based on their biological traits or behavior. It does not fear circumstances such as intrusion of privacy due to stolen smartcards or password being hacked. Biological traits such as voices, fingerprints, DNA, retinal and iris patterns are practically unique to each and every individual, which makes this method of proving one’s identity nearly impossible to falsify. Out of these various biometric traits, biometric verification through vein recognition had been explored and improvised over a sustained period of time, arising as a new form of authentication. The current vein identification method used for authentication consist of three modules: Vein Acquisition, Image Processing and Line Pattern Matching. Vein Acquisition is the process of capturing the blood vessels using a thermal scope. The acquired images were then processed by Near-Infrared imaging tool to pass into the next module. The final module, Line Pattern Matching, matches the processed vein images with the database available for authentication. In this project, we enhanced the current Image Processing module by introducing edge detection and connection methods to the Near-Infrared images to identify more traits of the veins. This project was done with some modification to a previously implemented biometric system which was used for palm print and hand vein recognition. C# and MATLAB were used in this project for edge detection. Visual Studio, an open source IDE, was explored but the results using the functions available for Visual Studio are not as desirable as MATLAB in terms of image processing.