Forest Fire Classifier – Instant detection of fire, smoke, and no-fire in images

Summary: Forest Fire Classifier is a web app that detects fire, smoke, fire+smoke, or no fire in forest images within approximately 500 ms. It uses server-side TensorFlow.js inference with Next.js and TypeScript, processing images in memory without storage to ensure privacy.

What it does

Users upload a forest photo to receive classification and confidence scores quickly. All image processing and inference run server-side using TensorFlow.js, with no images stored.

Who it's for

Developers and learners interested in production-level machine learning web apps built with Next.js, TypeScript, and TensorFlow.js.

Why it matters

It provides a privacy-focused, open-source example of deploying real-time image classification in a web app, demonstrating the transition from a college project to a functional product.