Fingerprint Based Students' Attendance System

ABSTRACT

Attendance is typically a list of people present at a particular event. Manual approach of attendance taking has not been very effective due to difficulty in keeping the attendance list over a long period of time, unnecessary time wastage during writing or signing, improper documentation, students forgetting to write or sign-the attendance paper among others. The aim of this project is to develop a technology assisted attendance system that will enhance efficient and effective means of taking attendance using fingerprint. Fingerprint was used as the unique physiological trait of an individual. This attendance system involved data acquisition of fingerprint samples of 44 students of the Federal University Oye - Ekiti. Image preprocessing techniques which includes image gray scale conversion, binarization, central line thinning of binarized image, dilation of thinned image and thinning of dilated image were carried out on the acquired data. The features of the preprocessed data were extracted using minutiae matching and classification of the extracted features was done using Euclidean distance. Evaluation was done using False Acceptance Rate (FAR), False Rejection Rate (FRR) and Accuracy. The attendance system was developed using MATLAB R2015a. Experiment was conducted with 168 sample images (both left and right thumb fingerprint) selected to train the system and 96 sample images (both left and right thumb fingerprint) to test the system's functionality. The system has a False Accept Rate of 0.83, False Reject Rate of0.01. It has an accuracy of 87% with average execution time of 4.81 secs. This project has developed a software prototype of a fingerprint based attendance system. The system can be improved by implementing it as a hardware prototype that uses real - time fingerprints from individuals during recognition as opposed to the currently implemented software prototype which uses offline images acquired earlier.