Development of a novel biometric authentication & identification of a finger print based attendance system using adaptive digital image processing (DIP) techniques

Abstract

This work  presents an finger print based attendance system as a Biometric Identification means, which was tested using a database of gray scale eye images in .bmp format.  Firstly, a basic type of LED scanner was used to capture fingerprint images. The images obtained are in gray scale. For a new employee the fingerprint is captured and is converted into integer format and is then encrypted using simple encryption technique and stored in the database. Three fixed points are considered on the image and template of size 15*15 starting from each of these 3 points is stored in the database.  Now, when a live fingerprint is scanned, it is converted into integer in similar fashion. The templates stored in the database are scanned on the input fingerprint and the system tries to match. The points, which give the best match, are considered. In this way, we get 3 patches and 3 points. Now, the relative distance and angles between each of the point for both the images are calculated. Their difference is calculated and a decision is made to determine if there is a match or not depending on a threshold error value.  Once a match is found, a signal is sent to the micro-controller, which indicates the match by glowing an LED for a short period of time (say 1 sec). At the same time the micro-controller accesses the real time clock to extract the exact date and time of the entry and exit of the employee. This is stored in its EPROM.  This work aims to give a new dimension to fingerprint-based attendance systems by incorporating the feature of storing and extracting the exact time of the identification, entry and exit of employees.