In-Browser Image Classifier

A lightweight, real-time image classification tool built with ml5.js (MobileNet) and p5.js. It performs Transfer Learning entirely in the browser, allowing users to train a custom model using their webcam without backend processing.

Try it here: redlightgreenlightml.netlify.app

How to Run

1. Use the Red/Blue buttons to capture ~15 labeled images for each category from your webcam.
2. Click Train Model to retrain the feature extractor.
3. Once trained, the app classifies the live video feed and updates the DOM dynamically.

Tech Stack

• ml5.js (Feature extraction & Transfer Learning)
• p5.js (DOM manipulation)
• Vanilla JS/HTML/CSS