SugarCare: Diabetes Companion

AI-powered diabetes management platform — Comprehensive prediction and detection system for diabetes care

A web application developed as a Major Project at the Institute of Engineering, Tribhuvan University. SugarCare combines machine learning and deep learning to assist in diabetes prediction, diabetic foot ulcer detection, and diabetic retinopathy screening.

---

Screenshots

---

Features

Diabetes Prediction System

• SVM-based prediction using clinical parameters:
  • Age, Gender, Hypertension, Heart Disease
  • Smoking Habits, BMI, HbA1c Level, Blood Glucose Level
• Accurate risk assessment for diabetes

Diabetic Foot Ulcer Detection

• CNN-powered image classification
• 94% accuracy in detecting foot ulcers
• Binary classification: Ulcer vs Non-ulcer

Diabetic Retinopathy Detection

• Advanced CNN model for retinal image analysis
• 80% accuracy in severity classification
• Stages: No DR → Mild → Moderate → Severe → Proliferative DR

Wellness Features

• 3D yoga exercise animations (Blender-rendered)
• Comprehensive diabetes management guidance

---

Tech Stack

Technology	Purpose
Python	Backend & ML algorithms
Django	Web framework
NumPy	Numerical computations
Pandas	Data preprocessing
Scikit-Learn	SVM implementation
TensorFlow	CNN training (foot ulcer)
PyTorch	CNN training (retinopathy)
Keras	Neural network design
OpenCV	Image preprocessing
PIL (Pillow)	Medical image handling
Bootstrap	Responsive UI design
Blender	3D yoga animations

---

Installation & Setup

1. Create Virtual Environment

# Install virtualenv
pip install virtualenv

# Create virtual environment
python -m venv hamro_environment

# Activate virtual environment
# Windows
hamro_environment\Scripts\activate
# Linux/macOS
source hamro_environment/bin/activate

2. Install Dependencies

pip install -r requirements.txt

3. Run the Application

# Start Django development server
python manage.py runserver

# Open in browser
# http://127.0.0.1:8000/

---

Model Performance

Model	Accuracy	Purpose
SVM	High	Diabetes risk prediction
CNN (Foot Ulcer)	94%	Ulcer detection
CNN (Retinopathy)	80%	Retinopathy staging

---

Documentation

Research Paper & Reports

• Research Paper
• Project Report

---

Team Members

Kathmandu Engineering College — BCT 077 Batch

• Deephang Thegim — @mrdeephang
• Dipesh Awasthi
• Esparsh Tamrakar
• Pankaj Karki

---

Project Context

Institution: Institute of Engineering, Tribhuvan University
College: Kathmandu Engineering College
Project Type: Major Project
Batch: BCT 077 (2077 Batch)
Version: 1.0.0

---

License

All Rights Reserved © 2025 SugarCare Team