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