ABSTRACT
•This project introduces “FingerDiab”, a real-time, non-invasive diabetes risk prediction system that leverages biometric fingerprint data and health features. Unlike conventional methods that rely on blood sampling and medical lab tests, this solution utilizes fingerprint image processing to extract physiological markers such as ridge count and fingerprint type (arch, loop, whorl), which are correlated with diabetes risk.
•A Random Forest Classifier trained on a labeled dataset combines this biometric input with additional parameters like Age, BMI, and Family History to assess risk levels—LOW, MEDIUM, or HIGH.
•Key Highlights:
•✅ Accuracy: 99% on test data
•✅ End-to-end web interface built with Flask
•✅ Fingerprint upload + real-time ML inference
•✅ No physical lab tests required
EXISTING SYSTEM
•The existing system, titled “A Genetic-based Framework for Diabetes Risk Prediction Using Parental Genetic Contributions“, utilizes inherited parental traits and clinical biomarkers to assess diabetes risk in children.
•Important health and genetic characteristics of the children were identified as significant contributors.
• This simple method shows promise in the integration of family factors with ML to early detect and accurately identify at risk cases, thereby carrying the potential to predict, based on inherited information from parents,
DISADVANTAGES
❌ Requires Genetic Data
• The system depends heavily on parental genetic markers, which are often not easily accessible for the general population—limiting large-scale applicability.
❌ Complex Preprocessing Pipeline
• Involves extensive data cleaning, imputation, feature encoding, and noise injection, which makes the model resource-intensive and hard to deploy in real time.
❌ High Computational Overhead
• The use of genetic algorithms, SMOTE, and multiple model training cycles across epochs increases training time and demands high-performance computing resources.
PROPOSED SYSTEM
•The proposed system, FingerDiab, is a web-based diabetes risk prediction platform that integrates biometric fingerprint analysis with user health data to assess the likelihood of diabetes in real time.
•Upon uploading a fingerprint image, the system processes it using image processing techniques to extract critical features such as ridge count and fingerprint pattern type (arch, loop, whorl).
• These features are combined with user-provided inputs like age, BMI, and family history to form a comprehensive health profile. A Random Forest classifier, trained on labeled biometric-health data, predicts the diabetes risk level—categorized as LOW, MEDIUM, or HIGH.
🔹 Real-Time Diabetes Risk Prediction
• Automatically analyzes fingerprint and health data to generate immediate risk assessments, enabling early detection and awareness.
🔹 Biometric Feature Extraction from Fingerprints
• Uses advanced image processing (OpenCV, PIL) to extract ridge count and determine fingerprint pattern type (arch, loop, whorl), eliminating the need for invasive procedures.
🔹 Multi-Feature Model Input
• Combines fingerprint-derived features with health attributes like Age, BMI, and Family History to ensure more accurate predictions.
🔹 High Accuracy Random Forest Model
• Achieves 99% prediction accuracy on test data, trained using labeled biometric-health datasets with categorical encoding and validation.
ADVANTAGES
- ✅ USER-FRIENDLY WEB INTERFACE
- ✅ INTERACTIVE DATA DASHBOARD
- ✅ SCALABLE & MODULAR ARCHITECTURE
- ✅ Reduces manual effort and data entry errors
SOFTWARE & TECH STACK
- Backend: Python with Flask or Django
- ML Frameworks: scikit-learn, TensorFlow/Keras
- Chatbot Engine: Dialogflow / Rasa or custom NLU
- Frontend: React.js or HTML/CSS/JS
- Database: MySQL or MongoDB
HARDWARE REQUIREMENTS
- Processor: Intel i3 / AMD Ryzen 3 or higher
- RAM: Minimum 4 GB (8 GB recommended for ML training)
- Storage: 100 GB
📽️ DEMO VIDEO
INCLUDED PACKAGE
- Base Paper
- Complete Source Code
- Complete Documentation
- Presentation Slides (PPT)
- Flow Diagram (UML)
- Database File
- Screenshots
- Execution Procedure
- ReadMe File
- Add‑ons & Supporting Softwares
- Video Tutorials
SUPPORT & SPECIALIZATION
- Support via Ticketing System
- Voice Conference Assistance
- Video On Demand for Setup & Training
- Remote Connectivity Support
- Code Customization on Request
- Document Customization Assistance
- Live Chat Support
CONCLUSION
•The proposed system, FingerDiab, presents an innovative and practical approach to early diabetes risk detection by combining biometric fingerprint analysis with machine learning.
• By utilizing non-invasive techniques and a highly accurate Random Forest model (99%), the platform enables real-time predictions through an intuitive web interface.
•This solution eliminates the need for traditional diagnostic procedures, making it ideal for preventive health screening in both urban and underserved areas.
• Its modular design, secure login system, and interactive dashboard make it user-friendly, scalable, and impactful.
•Ultimately, this project demonstrates how AI and biometrics can work together to create accessible, efficient, and intelligent healthcare tools, paving the way for smarter preventive care and greater health awareness.
📩 SUPPORT
For source code access, customization, or guidance, contact us at: xpertieee@gmail.com