RadiAid
Our Solution
We built an AI-powered mobile diagnostic system that analyzes standard X-rays to identify osteosarcoma. The tool runs entirely on smartphones, processes images in under 5 seconds, and currently achieves 93.75% accuracy.
Our solution has three core components:
Image Acquisition
AI/ML Model
User Interface
Physicians upload digital X-rays or capture photos of traditional films with a mobile device. We built a quality box attachment that standardizes smartphone capture using high CRI LED panels, anti-reflective surface, and fixed camera angle.
Our deep learning model is trained on annotated X-rays plus patient metadata (age, gender, lesion location). The attention condenser network focuses on diagnostically relevant regions, learning patterns that indicate malignancy.
Results show as clear probability scores and heatmaps, helping providers interpret outputs. Designed for non-specialists in rural areas to distinguish malignant tumors from benign conditions
The Process
1. Doctor photographs X-ray film using quality box
2. AI analyzes image in under 5 seconds
3. System displays probability score and highlights suspicious regions
4. Doctor makes informed referral decision