Open Cam Lab
A browser-based facial analysis experiment with real-time face detection, multiple bounding boxes, emotion labels, and image uploads.
The starting point
Real-time visual feedback needs to stay responsive while a model processes camera frames. The project explores how to connect that processing with a usable frontend.
What I built
I built the application with React, TypeScript, face-api.js, and Redux Toolkit. It supports both live detection and uploaded images, and was deployed on AWS Amplify.
Under the hood
The interface draws bounding boxes for multiple faces and updates predicted labels as the input changes. Those labels are model outputs rather than facts about a person. Redux Toolkit provides a shared place to manage application state.
What stayed with me
This project made frontend performance tangible: model processing and UI updates share the browser. It also gave me practical experience with state management and deploying a React application to AWS.