AuraMed: A Unified Multi-Pathology Diagnostic System

Breast Cancer Detection Result
PCOS Detection Result
Cervical Cancer Image Processing (bitmap)
Home Dashboard

Inspiration

AuraMed was inspired by the need for accessible and reliable healthcare diagnostics in underserved regions where specialist doctors are limited. We wanted to build an AI-powered platform that could assist in the early detection of Breast Cancer, Cervical Cancer, and PCOS while also making AI predictions transparent and trustworthy.

What it does

AuraMed is a unified multi-pathology diagnostic platform that analyzes medical images using Vision Transformers and Explainable AI. The system detects Breast Cancer, Cervical Cancer, and PCOS from medical scans and generates diagnostic reports with Grad-CAM heatmaps for visual interpretation.

How we built it

We built AuraMed using Python, PyTorch, TensorFlow, and FastAPI. The diagnostic engine is powered by the Vision Transformer (ViT-B/16) architecture with support from ResNet50 for feature extraction. We integrated Explainable AI using Grad-CAM, along with OpenCV and Matplotlib for visual heatmap generation. The platform follows a modular microservices-based architecture for scalable cloud deployment.

Challenges we ran into

One major challenge was handling medical imaging datasets with inconsistent quality and formats. Training Vision Transformers also required significant computational resources and optimization. Another challenge was implementing Explainable AI in a way that produced clinically meaningful visualizations without reducing prediction accuracy.

Accomplishments that we're proud of

We successfully created a working prototype capable of detecting multiple diseases within a single AI platform. Achieving strong diagnostic accuracy while integrating Explainable AI was a major milestone. We are also proud of developing a clean and user-friendly diagnostic dashboard for healthcare professionals.

What we learned

This project helped us gain practical experience in medical image processing, Vision Transformers, Explainable AI, cloud deployment, and scalable AI system design. We also learned how important interpretability and trust are in healthcare-focused AI systems.

What's next for AuraMed

We plan to expand AuraMed by adding support for additional diseases, improving cloud scalability, and integrating federated learning for privacy-preserving healthcare AI. We also aim to collaborate with healthcare institutions for real-world testing and clinical validation.