🤖 Baymax

Your Personal Health Companion

---

🩺 Baymax: AI-Powered Healthcare Appointment & Skin Disease Screening Platform

Baymax is a full-stack digital healthcare platform designed to streamline medical appointments and provide automated AI-driven skin disease screening. By combining a robust Django REST API backend, a modern React frontend, and a PyTorch-based Deep Learning model (ResNet-18), Baymax delivers intelligent doctor recommendations, appointment management, and multimodal skin lesion assessment.

---

🌟 Features

🔐 Patient & Doctor Portals

• Role-based authentication system for Patients and Doctors
• Custom onboarding workflows for each user type
• Detailed user profiles:
  • Patients: Age, Gender, Medical History
  • Doctors: Specialization, Availability Hours, Professional Bio
• JWT Authentication using SimpleJWT
• Secure access and refresh token workflow
• Automatic token management via Axios interceptors

---

🩺 Smart Doctor Suggestions

• Enter symptoms in plain text
• Intelligent symptom-to-specialization mapping
• Example recommendations:
  • Rash → Dermatologist
  • Cough → General Physician
  • Chest Pain → Cardiologist

---

📅 Appointment Management

Multi-Step Appointment Booking

Patients can:

1. Select a Doctor
2. Choose Appointment Date
3. Select Available Time Slot
4. Describe Symptoms

Appointment Dashboard

Track appointment status:

• Pending
• Confirmed
• Cancelled

Manage and review upcoming consultations easily.

---

🤖 AI-Powered Skin Disease Screening

Image-Based Disease Detection

Upload skin lesion images for classification using a fine-tuned ResNet-18 model trained on the HAM10000 Dataset.

Supported disease categories include:

• Actinic Keratoses
• Basal Cell Carcinoma
• Benign Keratosis-like Lesions
• Dermatofibroma
• Melanoma
• Melanocytic Nevi
• Vascular Lesions

Multimodal Prediction System

The prediction engine combines:

Image Features

• Uploaded skin lesion image

Clinical Metadata

• Itching
• Pain
• Burning sensation
• Duration
• Lesion location

This fusion improves prediction quality and diagnostic relevance.

Actionable Results

The system provides:

• Disease classification
• Confidence score
• Severity level
• Personalized care recommendations
• Triage warnings
• Doctor consultation recommendations for critical cases

---

🛠️ Technology Stack

Backend

Technology	Purpose
Django 5	Backend Framework
Django REST Framework	REST APIs
SimpleJWT	Authentication
django-cors-headers	Cross-Origin Support
SQLite	Database
Pillow	Image Processing
PyTorch	Model Inference

Frontend

Technology	Purpose
React 18	SPA Development
React Router v6	Routing
Tailwind CSS	Styling
Axios	API Communication
Radix UI	UI Components
Lucide React	Icons

Machine Learning

Technology	Purpose
PyTorch 2.0	Deep Learning
torchvision	Image Processing
scikit-learn	Data Splitting
Pandas	Data Handling
NumPy	Numerical Computation
TensorBoard	Training Visualization

---

📂 Project Structure

baymax/
├── appointments/
│   ├── models.py
│   ├── serializers.py
│   └── views.py
│
├── baby/
│   ├── models.py
│   ├── serializers.py
│   └── views.py
│
├── baymax/
│   ├── settings.py
│   ├── urls.py
│   └── wsgi.py
│
├── diagnosis/
│   ├── views.py
│   └── ml_inference.py
│
├── frontend/
│   ├── public/
│   └── src/
│       ├── components/
│       ├── Api.js
│       └── App.js
│
├── ml_training/
│   ├── outputs/
│   ├── config.py
│   ├── train.py
│   ├── train_multimodal.py
│   ├── inference.py
│   └── QUICKSTART.md
│
├── db.sqlite3
├── manage.py
└── README.md

---

🚀 Getting Started

Prerequisites

• Python 3.10+
• Node.js 18+
• npm

---

1️⃣ Backend Setup

Create Virtual Environment

Windows

python -m venv venv
venv\Scripts\activate

Linux / macOS

python3 -m venv venv
source venv/bin/activate

Install Dependencies

pip install django djangorestframework djangorestframework-simplejwt django-cors-headers pillow torch torchvision pandas numpy scikit-learn tqdm matplotlib tensorboard

Apply Migrations

python manage.py migrate

Run Backend Server

python manage.py runserver

Backend URL:

http://127.0.0.1:8000/

---

2️⃣ Frontend Setup

Navigate to frontend directory:

cd frontend

Install packages:

npm install

Start development server:

npm start

Frontend URL:

http://localhost:3000/

---

3️⃣ Machine Learning Setup (Optional)

The diagnosis system requires trained model weights:

resnet18_best.pth
or
resnet18_multimodal_best.pth

stored inside:

ml_training/outputs/

Dataset Structure

archive/
├── HAM10000_metadata.csv
└── processed_images_dataset/
    └── processed_images/

Train Image Classification Model

cd ml_training
python train.py

Train Multimodal Model

python train_multimodal.py --csv path/to/metadata.csv --image-root path/to/images

Run Inference

python inference.py --image "path/to/lesion_image.jpg"

---

🔌 API Endpoints

Method	Endpoint	Description	Authentication
POST	/api/users/login/	Obtain JWT Access & Refresh Tokens	No
POST	/api/users/token/refresh/	Refresh Access Token	No
GET	/api/users/profile/	Fetch User Profile	Yes
POST	/api/users/register/	General Registration	No
POST	/api/users/register/patient/	Patient Registration	No
POST	/api/users/register/doctor/	Doctor Registration	No
GET	/api/users/doctors/	Fetch All Doctors	No
GET	/api/users/suggest-doctors/	Symptom-Based Recommendations	No
GET	/api/appointments/	Retrieve Appointments	Yes
POST	/api/appointments/	Book Appointment	Yes
PUT/PATCH	/api/appointments/{id}/	Update Appointment	Yes
DELETE	/api/appointments/{id}/	Cancel Appointment	Yes
POST	/api/diagnosis/skin-check/	AI Skin Disease Screening	No

---

🧠 Machine Learning Pipeline

Dataset

• HAM10000 Skin Lesion Dataset
• Stratified Train / Validation / Test Split (70/15/15)

Model

• ResNet-18 Transfer Learning
• Fine-tuned for 7 skin disease classes
• Class-weighted loss handling
• TensorBoard training monitoring

Multimodal Fusion

Combines:

• CNN Image Embeddings
• Patient Clinical Metadata

to generate enhanced disease predictions.

---

⚕️ Medical Disclaimer

Baymax is intended for educational, research, and demonstration purposes only.

The AI predictions, recommendations, and triage advice generated by the system are not a substitute for professional medical diagnosis, treatment, or clinical judgment.

Always consult a qualified dermatologist or healthcare professional regarding any medical condition or health concern.

---

👨‍💻 Author

Developed using:

• Django REST Framework
• React
• PyTorch
• Tailwind CSS
• JWT Authentication
• HAM10000 Dataset

---

📜 License

This project is intended for educational and research purposes.

Feel free to fork, modify, and extend the project for learning and experimentation.