EMMANUEL_ARO_/>← Back to work

// CASE STUDY

Real-Time Computer Vision Attendance Platform

Engineered a high-speed biometric attendance system integrating webcam image streaming with OpenCV computer-vision pipelines. Automated real-time facial capture, normalization, and identification against a secured relational store.

Role
Full-Stack Developer
Year
2023
Context
Computer Vision
Status
Archived
PHPJavaScriptOpenCVMySQL
01.

Overview

This biometric platform replaces manual roll calls and keycard systems with a real-time, computer-vision driven attendance tracker powered by OpenCV.

02.

The problem

Traditional attendance tracking methods are slow, prone to buddy-punching, and require physical contact or specialized hardware. Organizations need frictionless, secure, and instantaneous verification.

03.

My approach

I developed a web-based capture interface using JavaScript that streams webcam data directly into an OpenCV processing pipeline. The backend rapidly normalizes the incoming frames and performs facial identification against a secure MySQL relational database.

04.

Architecture decisions

Real-time normalization — Built a preprocessing layer that handles varied lighting conditions and facial angles before attempting classification, drastically reducing false negatives.

Hybrid stack — Paired a lightweight PHP/MySQL backend for rapid data storage with a heavy OpenCV processing layer, balancing web accessibility with intensive mathematical computations.

05.

Outcome

The platform achieved highly reliable facial identification in real-time, effectively demonstrating the viability of browser-based biometric access control without requiring proprietary hardware.

View code ← All projects