ISSN : 2663-2187

A ROBUST IOT-BASED FACE RECOGNITION ATTENDANCE SYSTEM WITH LIVENESS DETECTION AND ADAPTIVE LOW-LIGHT ENHANCEMENT

Main Article Content

Mr. Anirban Majumder,Prof (Dr.) Jhunu Debbarmab , Prof (Dr.) Gitanjali Sahac, Prof (Dr.) Shimal Das
» doi: 10.48047/AFJBS.8.10.2026.51-73

Abstract

The main goal of this study was to create an IoT-based face recognition attendance system to improve security, automation and performance in real time environments. This system combines face recognition with liveness detection methods such as blink detection and head movement verification to reduce spoofing attempts using printed photos and static images. To improve performance in low light, we used contrast-limited adaptive histogram equalization (CLAHE) as a pre-processing method. These improvements boost facial exposure and increase the recognition accuracy in dark conditions. The system integrates an ESP8266 IoT module and a PIR motion sensor to automate the recognition process. Instead of continuously processing the video frames, our system activates only when motion is detected which reduces the unnecessary computation and improves the energy use. Once activated the system performs real-time face recognition, automatically records attendance and prevents duplicate entries through anti-spam method. According to experimental observations the system functions dependably in a variety of real-world scenarios and lighting circumstances. The suggested method provides an automatic attendance management system for companies and schools that is safe, scalable and affordable.

Article Details