AI Skin Cancer Detection System with 83% Accuracy DeployedAI Skin Cancer Detection System with 83% Accuracy Deployed
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Back to feedPost
osama's avatar
osama zaman
• Mar 15
Built a full end-to-end AI skin cancer detection system — live and deployed.
Here's what I shipped:
→ Fine-tuned EfficientNet-B0 on 10,015 real dermoscopy images (HAM10000 dataset) → 83.3% accuracy across 7 disease classes including Melanoma and Basal Cell Carcinoma → FastAPI backend serving predictions in under 80ms → Clean clinical frontend — upload a photo, get instant AI analysis → Fully deployed: GitHub Pages + Hugging Face Spaces
Live demo: https://usamazaman123.github.io/dermai/frontend Source code: https://github.com/usamazaman123/dermai
Tech stack: • PyTorch + timm (EfficientNet-B0) • FastAPI + Uvicorn • Trained on Kaggle GPU P100 • Pure HTML/CSS/JS frontend — no framework
Post image
Post image
Back to feed
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started