Image Classifier

Image Classifier is a polished browser-based vision demo that analyzes uploaded images directly on the client side. It estimates a best-fit category, surfaces confidence scores, breaks down visual signals like brightness and contrast, and stores recent analyses in local history.

Live Demo

• Portfolio route: https://rabin-portfolio-rust.vercel.app/image-classifier/
• Portfolio homepage: https://rabin-portfolio-rust.vercel.app/
• Repository: https://github.com/rabinregmi188/Image-Classifier

Highlights

• Drag-and-drop upload flow with instant preview
• Client-side image analysis using the Canvas API
• Top-3 prediction scoring with animated confidence bars
• Signal breakdown cards for brightness, contrast, saturation, and scene bias
• Local history saved with localStorage
• Responsive dashboard with colorful glassmorphism styling

Stack

• JavaScript
• HTML
• CSS
• Canvas API
• LocalStorage

Run locally

Open public/index.html in a browser, or serve the folder with a simple static server.

Project Notes

This version is intentionally static so it can deploy cleanly inside the portfolio, the same way the other live project routes work.