ImageMind.

Overview

A deep CNN image classifier trained on CIFAR-10, deployed as a Streamlit web app for real-time image classification with confidence scores and top-3 predictions.

What's inside

• Real-time image classification across 10 CIFAR-10 categories
• Confidence scores and top-3 predictions display
• Custom CNN architecture with data augmentation pipeline
• Modern, responsive Streamlit UI
• Mobile-friendly design for on-the-go inference

Challenges

Optimising the CNN architecture to achieve strong accuracy on CIFAR-10 while keeping inference fast enough for a real-time web demo was a key trade-off throughout development.

What I took away

Strengthened my end-to-end ML pipeline skills — from model training and evaluation through to deploying an interactive inference app with Streamlit.