HealthAI

📖 Description

The HealthAI is an innovative web application designed to enhance healthcare accessibility through advanced AI technologies. It offers features such as multi-lingual medical query support along with precription analysis, x-ray analysis, and a symptom-to-disease prediction tool. The system leverages the Intel GENAI toolkit and integrates machine learning models for efficient, accurate healthcare assistance. With user authentication and a personalized experience, this tool aims to revolutionize how individuals seek medical advice and information.

---

📚 Table of Contents

1. Features
2. Technologies Used
3. Website
4. Optimizations
5. Acknowledgments
6. Future Work
7. Screenshots/Demo

---

✨ Features

Here are the features of HealthAI with even more concise descriptions:

• Multi-lingual Medical Query Support: Users can input queries in multiple languages and receive accurate responses through text and voice inputs.

• Fracture Detection and Prediction: HealthAI uses YOLOv8 to detect and predict fractures from medical images quickly and accurately.

• Symptom-to-Disease Prediction Tool: Users select symptoms, and the tool predicts potential illnesses using classifiers like SVC and Random Forest.

• Prescription Analysis: Tesseract and Llama 3.1 analyze prescriptions, extracting details and answering medication-related queries.

• User Authentication: A secure system allows users to create accounts and store medical inquiries, ensuring personalized experiences and data privacy.

---

⚙️ Technologies Used

Here’s the Technologies Used section for HealthAI, highlighting the key technologies and tools implemented in the project:

• Intel GENAI Toolkit: Utilized for developing AI models and enhancing performance in medical applications.

• YOLOv8: Employed for efficient fracture detection and prediction from medical images.

• Llama 3.1: Integrated for natural language processing and prescription analysis tasks.

• Tesseract: Used for optical character recognition (OCR) to extract text from prescription images.

• Scikit-learn: Leveraged for implementing machine learning algorithms like SVC, Random Forest, and Naive Bayes for symptom prediction.

• Firebase: Implemented for secure user authentication and data storage, ensuring scalability and privacy.

• HTML/CSS/JavaScript/node.js: Core technologies for building the web application’s user interface and interactivity.

• Flask: Utilized as the backend framework for handling API requests and serving the web application.

---

🌐 Website

The HealthAI web application serves as an accessible platform for users to interact with various healthcare features seamlessly. Designed with a user-friendly interface, it allows individuals to input medical queries, analyze symptoms, detect fractures, and gain insights into their prescriptions.

• Live Demo: [https://nodejs-server-wc8r.onrender.com/]
• Source Code: [https://github.com/LokeshYarramallu/Project_oneAPI_hack_kpr]
• Youtube: [https://youtu.be/oVsafKuqhlM]

Getting Started

To explore the application:

1. Visit the website link.
2. Create an account or log in to access personalized features.
3. Use the navigation to explore various tools and functionalities.

---

⚡ Optimizations

The Intel GENAI Toolkit significantly enhances the performance of the HealthAI application through several key optimizations:

• Faster Model Training: The toolkit utilizes optimized algorithms and Intel's multi-core processing capabilities, enabling quicker training of machine learning models and improved accuracy.

• Real-time Inference: With techniques like model quantization and pruning, the toolkit reduces model size while maintaining accuracy, allowing for rapid responses to user queries.

• Multi-threading Support: By enabling multi-threaded execution, HealthAI can perform multiple operations simultaneously, reducing wait times for complex tasks like image detection and query processing.

• OpenVINO Integration: The toolkit integrates with Intel OpenVINO, optimizing deep learning models for efficient deployment, particularly enhancing the speed of fracture detection with YOLOv8.

• Performance Monitoring: The toolkit includes tools for profiling and monitoring performance, helping developers identify bottlenecks and optimize resource usage effectively.

These optimizations ensure that HealthAI delivers fast, reliable, and accurate healthcare assistance to users.

---

🙏 Acknowledgments

• Intel for the GENAI Toolkit and OpenVINO.

• Kaggle for the bone fracture detection dataset here and dataset for symptom to disease training data.

• GitHub for version control.

• Open Source Community including Llama and Groq for essential libraries and frameworks.

• Team Members for their hard work and collaboration:

  Lokesh Yarramallu

  

  Tarun Sudheer

  

---

🚀 Future Work

• Enhanced User Interface: Improve the UI/UX for better accessibility and user engagement.
• Additional Languages: Expand multi-lingual support to cover more languages for broader accessibility.
• Integration with Telemedicine: Develop features for virtual consultations and remote monitoring.
• Advanced Diagnostics: Incorporate more sophisticated diagnostic tools and algorithms for improved accuracy.
• Mobile Application: Create a mobile version of the application for on-the-go access to healthcare tools.
• User Feedback Mechanism: Implement a feedback system to gather user insights for continuous improvement.

---

📸 Screenshots/Demo

ARCHITECHTURE

Visuals of the application in action to give users a better understanding of the interface and functionalities.

1. Home Page Overview:
   
   The home page of HealthAI, featuring a clean, user-friendly interface for navigating the main features, including medical queries, symptom detection, and fracture analysis.

2. Medical Query with Multi-Lingual Support:
   
   A demonstration of the system handling both text and voice input for medical queries in multiple languages, using Llama 3.1.

3. Fracture Detection and Prediction:
   Example of the fracture detection interface, where the YOLOv8 model predicts fractures based on user-uploaded X-ray images.

4. Symptom to Disease Prediction Tool:
   
   Interface showing the tool where users input symptoms, and the system predicts possible diseases using SVC, Random Forest, and Naive Bayes classifiers.

5. Prescription Analysis Tool:
   Demonstration of the prescription analysis feature, which extracts and analyzes text from prescriptions using Tesseract and Llama 3.1.

Usage

To Use this project open your terminal and enter the following commands

  cd backend/
  npm install
  node app.js

In other terminal

  cd backend/models
  pip install -r req.txt
  python app.py

API and Firebase Configuration Setup

To properly configure your project, follow the steps below:

1. Replace <YOUR_GROQ_API>

• Visit Groq Console.
• Generate a new API key.
• Replace <YOUR_GROQ_API> in the code with your newly generated API key.

2. Replace <YOUR_FIREBASE_CONFIG>

• Go to Firebase Console.
• Navigate to your project and select Project Settings.
• Scroll down to the Firebase SDK snippet section and copy your Firebase configuration.
• Replace <YOUR_FIREBASE_CONFIG> in the code with the copied configuration.

🚀 Closing Thoughts

HealthAI is more than just a tool—it’s a step toward making healthcare more accessible, efficient, and intelligent. By leveraging cutting-edge technologies like Intel's GENAI Toolkit and OpenVINO, we've pushed the boundaries of what's possible in medical assistance. This is just the beginning.

We believe in the power of collaboration, innovation, and the endless potential of AI to transform lives. HealthAI stands as a testament to that belief, and we’re excited to see where it evolves from here.

Together, we are building the future of healthcare—one line of code at a time.