
HummingBird : Real-Time Video Captioning Web Application
May 2025 - Present
A real-time video captioning web app that generates live textual descriptions from webcam input. Designed to enhance accessibility for the hearing impaired and improve media content analysis, it leverages open-source models and WebSockets to deliver accurate, low-latency captions in real time via a user-friendly interface.
Technologies Used
React.js
WebSockets
Python (Flask/Django)
PyTorch
Tailwind CSS