BioSense AI predicts early-stage diabetes risk by combining patient vitals with real-time facial emotion analysis — giving clinicians a richer picture than biomarkers alone.
Today, 22 Jun 2026
How it works
The patient submits vitals through the portal and optionally captures a webcam frame. Both signals are processed in parallel and fused into a single clinical recommendation.
Patients enter glucose, BMI, blood pressure, age, and other biomarkers. Missing values are imputed automatically using KNNImputer.
An optional webcam snapshot is passed through the FER library. The detected affect (e.g. stressed, neutral) is translated into a clinical descriptor.
The Clinical Fusion Engine combines the Random Forest risk probability, SVM confidence score, and emotion signal into one recommendation with SHAP explainability.
Built with
Deployed on Railway (backend) and Vercel (frontend).
Team
Vardhaman College of Engineering, Hyderabad — B.Tech CSE, 2025–26.
Syed Uzair Mohiuddin
Full Stack & AI Engineer
Frontend, FastAPI backend, deployment pipeline, and multimodal inference integration.
Sarasam Chinmaee Reddy
AI Architect
System architecture, clinical AI workflow design, and ML pipeline strategy.
| Member | Title | Scope |
|---|---|---|
| Syed Uzair Mohiuddin | Full Stack & AI Engineer | Next.js frontend, FastAPI backend, REST API, Railway/Vercel deployment, auth, UI/UX |
| Sarasam Chinmaee Reddy | AI Architect | System design, multimodal AI workflow, clinical pipeline architecture, research |
| Manohar Yadav Boddu | ML Engineer | Model training, preprocessing, feature engineering, SHAP/PDP explainability, evaluation |