DermaDetectAI
A simplified and streamlined version of the original DermaDetectAI project. Built with Streamlit and powered by PyTorch, this app enables quick and accurate skin disease detection.
Executive Summary
Providing a quick, accessible, and lightweight tool for preliminary skin disease identification directly in the browser.
DermaDetectAI-v2 is an AI-powered diagnostic tool rebuilt for efficiency. Developed as a college initiative by a team of four, it leverages PyTorch and Streamlit to classify skin images into multiple disease categories. The system is fully web-based, ensuring that users can receive preliminary health insights without the need for complex software installations.
Core Infrastructure
Fast, Lightweight UI
Built with Streamlit for quick interaction and easy deployment.
Deep Learning Powered
Uses a PyTorch model trained on a diverse skin disease dataset.
Instant Image Upload
Upload your skin image and get predictions instantly.
Fine-tuned Classification
Optimized inference pipeline for accurate results.
Design Philosophy
Rebuilt as an upgraded, simplified version of the original DermaDetectAI for a college initiative to ensure high efficiency and ease of use for general users.
Successfully deploying a complex PyTorch model on a lightweight Streamlit interface, allowing for sub-second inference in a web-based environment.
Technical Architecture
Optimizing the inference pipeline to ensure the model weights load quickly in a serverless Streamlit environment without sacrificing accuracy.
Engineered With
- Streamlit
- PyTorch
- TorchVision
- Pillow
- Pandas
Performance Goal
- Instant inference results
- Low memory footprint for web deployment
- Seamless image processing pipeline
System Integrity
- Secure image handling
- Robust model weight loading
- Reliable classification accuracy
