Skip to content

feruz-bb/cardioscan.github.io

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 

Repository files navigation

CardioScan 🫀

AI-Powered Cardiovascular Disease Early Detection System

A conceptual demo platform that leverages machine learning to assess a person's 5-year cardiovascular risk from basic biometric data — designed for the Uzbekistan healthcare context.


The Problem

Cardiovascular diseases (CVD) are the leading cause of death in Uzbekistan, accounting for 57.2% of all recorded deaths in the first 9 months of 2024. Key systemic challenges include:

  • Only 29 cardiologists per million people — far below European standards
  • CVD is increasingly affecting younger age groups (16–18)
  • Limited access to high-tech diagnostic equipment in regional areas
  • No scalable early-warning tool for primary care settings

The Solution

CardioScan is an AI-based risk assessment platform that takes 8–12 basic patient inputs (age, smoking status, blood pressure, lab results, etc.) and outputs a 0–100 Cardiovascular Risk Index for the next 5 years.

Core Features

Feature Description
Risk Prediction Gradient Boosting model estimates 5-year CVD risk from routine clinical data
Explainable AI (XAI) SHAP-based feature importance shows why the model gave a certain score
Personalized Recommendations Gemini API generates lifestyle, diet, and exercise advice tailored to the risk level
Triage Support Flags high-risk individuals for urgent cardiology referral

Tech Stack

Layer Technology
Frontend React, Tailwind CSS
Backend Python (Flask / Django)
Database Google Cloud Firestore
ML Model Scikit-learn, Pandas — Gradient Boosting Classifier
Generative AI Gemini API
Explainability SHAP

How It Works

Patient inputs basic data
        ↓
Gradient Boosting model generates Risk Index (0–100)
        ↓
SHAP explains the top contributing risk factors
        ↓
Gemini API generates personalized health recommendations
        ↓
High-risk patients are flagged for cardiology referral

Demo Output Example

Risk Index: 78% — HIGH RISK GROUP

Top contributing factors:
  Age (55):            ↑ 45% influence
  Current smoker:      ↑ 30% influence
  Blood pressure 145/95: ↑ 15% influence

AI Recommendation:
  "We recommend quitting smoking immediately and engaging in at least
   150 minutes of brisk walking per week. This could significantly
   reduce your cardiovascular risk."

Roadmap

Phase 1 (Current) — Concept & Demo
  └─ Problem definition, PoC, AI simulation demo

Phase 2 — First AI Model & Testing
  └─ Dataset collection, model training, initial evaluation

Phase 3 — MVP Platform
  └─ Backend API + React frontend integration
  └─ User auth + Firestore result storage

Phase 4 — Pilot & Clinical Validation
  └─ Validation with local cardiologists
  └─ Pilot deployment in medical institutions

Project Status

⚠️ This is a conceptual demo built for the AI500 competition. The risk scores shown are simulated and not intended for real medical use.


Author

Baqoyev Feruzbek — Solo Developer Skills: Python (AI/Backend), React (Frontend)


Motivation

CVD-related deaths in Uzbekistan are the primary social driver behind this project. The goal is to build something genuinely useful and deployable — not just a demo, but a foundation for a real clinical tool.


CardioScan — Advancing Preventive Medicine with Innovative Solutions Built for the AI500 Competition · 2025

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages