HemoGraph: AI-Powered Anemia Analysis

Inspiration

Have you ever seen someone who constantly feels weak or dizzy... and later finds out it was because of anemia — something so simple, yet so common?

That moment made me think: What if an app could detect anemia instantly from a blood report and help people understand it clearly?

Hi everyone! I'm Thrishanth Reddy, a first-year B.Tech student from NIT Warangal, and this is my project — HemoGraph — an AI-powered anemia analysis. As I honestly mentioned in my presentation, I've just started learning how to code! But I had an idea, a purpose, and the curiosity to learn.

I didn't have years of experience — but I had powerful AI tools and the determination to build something meaningful.

What it does

HemoGraph is an AI-powered anemia detection and analysis app that:

• Analyzes blood reports (manual entry or OCR upload)
• Detects anemia type using rule-based AI
• Provides confidence scores and explanations
• Stores patient history and tracks progress
• Generates personalized recovery plans with diet tips
• Features an educational section on anemia
• Works offline with a beautiful, calming interface

Unlike other apps that just say "you might have anemia," HemoGraph provides differential diagnosis — it tells you what type, how severe, and why.

How we built it

As a beginner who just started coding, I built this entire app using cutting-edge AI tools:

Frontend Development:

• DeepSeek - Built the complete frontend interface and user interactions
• Generated React components with Tailwind CSS styling
• Created the ECG-style animations and theme toggles

Backend Development:

• DeepSeek - Implemented core logic and data handling
• Gemini - Developed the anemia detection algorithm and rule-based AI model
• Built patient data storage using LocalStorage

UI/UX Design:

• ChatGPT - Designed the visual layout and user experience flow
• Created the calming, medical-grade interface concept
• Designed form layouts and navigation structure

Research & Refinement:

• Perplexity - Deep research on anemia types, blood parameters, and medical accuracy
• Refined feature specifications and implementation details
• Validated medical information and detection logic

Development Approach: I used these AI tools as my coding mentors, learning as I built. Each tool helped me understand different aspects:

• DeepSeek taught me React and component architecture
• Gemini helped me understand AI logic and algorithms
• ChatGPT guided me on design principles
• Perplexity ensured medical accuracy

This project proves that with the right AI tools and determination, even beginners can build impactful, production-quality applications.

Challenges we ran into

• Learning to effectively communicate with AI tools to get desired outputs
• Understanding React and JavaScript concepts from scratch
• Building an explainable AI model that shows reasoning, not just results
• Ensuring medical accuracy while implementing the detection algorithm
• Designing a medical-grade interface that feels professional yet accessible
• Implementing OCR functionality (currently in demo stage)
• Balancing feature richness with offline capability
• Debugging issues when AI-generated code didn't work perfectly

What's next for HemoGraph

• Hospital-grade OCR: Full pipeline to understand any report format
• Cloud AI integration: More accurate ML models for detection
• Multilingual support: Make it accessible to everyone
• Clinician feedback: Medical professional validation
• Mobile app: Native iOS and Android versions
• Production-ready: Transform from demo to real-world medical assistant
• API integration: Connect with healthcare systems
• Advanced analytics: Track population-level anemia trends

Built With

• ai
• chatgpt
• deepseek
• gemini
• perplexity