Overview
A prototype for a software-as-a-service platform that provides live visual analytics on video feeds. Users register, connect video feeds, and receive real-time object detection and counting — initially focused on pedestrian and vehicle analytics.
Built as an undergraduate capstone project (in lieu of thesis), this demonstrated end-to-end full-stack development: user authentication, video stream processing, ML inference, and a responsive dashboard UI.
Features
Live object detection
Supports pretrained or fine-tuned detection models running inference on live video feeds in real time.
User authentication
Registration and login system to manage access to video feeds and analytics dashboards.
Flask backend
Python-powered backend handling video stream processing, model inference, and API endpoints.
Responsive UI
CSS and Bootstrap-based dashboard for viewing live feeds alongside analytics metrics.
Built with
Python
Flask
Object Detection
OpenCV
Bootstrap
CSS