Iris Detection for Attendance Monitoring in Educational Institutes Amidst a Pandemic: A Machine Learning Approach

Abstract

Amid the COVID-19 pandemic, the imperative for alternative biometric attendance systems has arisen. Traditionally, fingerprint and facial recognition have been employed; however, these methods posed challenges in adherence to Standard Operational Procedures (SOPs) set during the pandemic. In response to these limitations, iris detection has been advanced as a superior alternative. This research introduces a novel machine learning approach to iris detection, tailored specifically for educational environments. Addressing the restrictions posed by COVID-19 SOPs, which permitted only 50% of student occupancy, an automated e-attendance mechanism has been proposed. The methodology comprises four distinct phases: initial registration of the student's iris, subsequent identity verification upon institutional entry, evaluation of individual attendance during examinations to assess exam eligibility, and the maintenance of a defaulter list. To validate the efficiency and accuracy of the proposed system, a series of experiments were conducted. Results indicate that the proposed system exhibits remarkable accuracy in comparison to conventional methods. Furthermore, a desktop application was developed to facilitate real-time iris detection.