Back to Projects
Mobile

Sanjeevani

Sanjeevani digitizes healthcare records and provides intelligent medical insights using OCR for prescription scanning, NLP for symptom analysis, and ML models for preliminary health assessments.

FlutterPythonSupabaseMachine Learning

System Architecture Diagram

Overview

An AI-powered healthcare platform that digitizes medical records and provides intelligent health insights through OCR, NLP, and machine learning.

Problem Statement

Paper-based health records are easily lost, difficult to share between doctors, and impossible to analyze for patterns. Patients lack tools to understand their health data.

Research

Studied HIPAA-equivalent Indian health data regulations (DISHA). Researched OCR accuracy for handwritten prescriptions and NLP models for medical text understanding.

Architecture

Flutter for cross-platform mobile, Python backend for ML/AI processing, Supabase for secure health data storage with Row Level Security.

Technical Decisions

Chose Flutter over React Native for better performance on low-end Android devices prevalent in India. Used Python for ML backend due to ecosystem maturity.

Tradeoffs

Prioritized OCR accuracy over processing speed. Accepted longer scan times for more reliable text extraction from handwritten prescriptions.

Challenges

Handling diverse handwriting styles in prescriptions, ensuring medical data privacy, and building accurate symptom-to-condition mapping.

Lessons Learned

Healthcare AI needs extremely conservative confidence thresholds. A wrong medical suggestion is worse than no suggestion at all.

Future Roadmap

Telemedicine integration, pharmacy network for prescription fulfillment, and family health dashboard.