YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
🩺 Chest X-Ray Pneumonia Detection using Deep Learning
📌 Project Overview
This project is an AI-powered Chest X-Ray Pneumonia Detection System that automatically classifies chest X-ray images into:
- ✅ NORMAL
- 🦠 PNEUMONIA
The project demonstrates multiple Deep Learning approaches and compares their performance using:
- Custom CNN (10 Epochs)
- Improved CNN (20 Epochs)
- Transfer Learning (VGG16)
The trained models are deployed using Flask, while the frontend is developed using Streamlit, creating a complete end-to-end AI medical imaging application.
🚀 Features
✔ Binary Classification
✔ Three Deep Learning Models
✔ Custom CNN Architecture
✔ Improved CNN Architecture
✔ Transfer Learning using VGG16
✔ Real-time Image Prediction
✔ Flask REST Backend
✔ Streamlit Interactive UI
✔ Hugging Face Model Hosting
✔ GPU Training Support
✔ Model Comparison
✔ Confidence Score Prediction
✔ Production Ready
🧠 Problem Statement
Pneumonia is one of the leading causes of death worldwide.
Radiologists inspect Chest X-rays manually which is:
- Time Consuming
- Error Prone
- Expensive
This project automates the diagnosis process using Deep Learning.
📂 Dataset
Chest X-Ray Dataset
Dataset
│
├── train
│ ├── NORMAL
│ └── PNEUMONIA
│
├── validation
│ ├── NORMAL
│ └── PNEUMONIA
│
└── test
├── NORMAL
└── PNEUMONIA
Image Format
- JPG
- JPEG
- PNG
Classes
| Label | Description |
|---|---|
| NORMAL | Healthy Lung |
| PNEUMONIA | Infected Lung |
⚙️ Tech Stack
Programming
- Python
Deep Learning
- TensorFlow
- Keras
Computer Vision
- OpenCV
- NumPy
- Matplotlib
Backend
- Flask
Frontend
- Streamlit
Deployment
- Hugging Face
- Render
Version Control
- Git
- GitHub
🏗 Complete Project Architecture
Chest X-Ray Image
│
▼
Upload Image (UI)
│
▼
Streamlit Frontend
│
▼
Flask REST API
│
▼
Image Preprocessing
│
┌──────────────────┼───────────────────┐
│ │ │
▼ ▼ ▼
CNN Model CNN 20 Model VGG16 Model
│ │ │
└──────────────────┼───────────────────┘
│
▼
Prediction Probability
│
▼
NORMAL / PNEUMONIA Result
│
▼
Display Prediction
📊 End-to-End Workflow
Dataset
│
▼
Image Loading
│
▼
Image Preprocessing
│
▼
Resize Images (100x100)
│
▼
Convert to Array
│
▼
Normalize Images
│
▼
Train / Validation Split
│
▼
Model Training
│
▼
Model Evaluation
│
▼
Save Best Model
│
▼
Deploy Model
│
▼
User Upload Image
│
▼
Prediction
🧹 Image Preprocessing Pipeline
Each X-Ray undergoes the following preprocessing steps:
Step 1
Load Image
↓
Step 2
Convert to Grayscale (CNN Models)
↓
Step 3
Convert to RGB (VGG16)
↓
Step 4
Resize
100 × 100
↓
Step 5
Convert to NumPy Array
↓
Step 6
Normalize Pixel Values
0 → 255
↓
0 → 1
↓
Step 7
Feed into Model
🤖 Model 1
Custom CNN (10 Epochs)
Architecture
Input (100×100×1)
↓
Conv2D (64)
↓
MaxPooling
↓
Dropout
↓
Conv2D (128)
↓
MaxPooling
↓
Dropout
↓
Conv2D (256)
↓
MaxPooling
↓
Dropout
↓
Flatten
↓
Dense (64)
↓
Dropout
↓
Dense (1)
↓
Sigmoid
Loss
Binary Crossentropy
Optimizer
Adam
Epochs
10
🤖 Model 2
Improved CNN (20 Epochs)
Architecture
Input
↓
Conv2D (64)
↓
ReLU
↓
MaxPooling
↓
Dropout
↓
Conv2D (128)
↓
ReLU
↓
MaxPooling
↓
Dropout
↓
Conv2D (256)
↓
ReLU
↓
MaxPooling
↓
Dropout
↓
Flatten
↓
Dense (64)
↓
Dropout
↓
Dense (1)
↓
Sigmoid
Epochs
20
Optimizer
Adam
Loss
Binary Crossentropy
🤖 Model 3
Transfer Learning (VGG16)
Pretrained
ImageNet
Frozen Layers
All VGG16 Convolution Layers
Custom Head
Flatten
↓
Dense (256)
↓
Dense (128)
↓
Dense (64)
↓
Dense (1)
↓
Sigmoid
Callbacks
- ModelCheckpoint
- EarlyStopping
🧠 Why VGG16?
Instead of training from scratch,
VGG16 already knows how to detect
- Edges
- Shapes
- Textures
- Patterns
Only the classifier is trained on Chest X-rays.
This greatly improves performance while reducing training time.
📁 Saved Models
model_xray.h5
Custom CNN
model_pre.h5
Improved CNN
best_model.keras
Best Transfer Learning Model
📈 Training Strategy
- GPU Training
- Batch Size = 4 (CNN)
- Batch Size = 32 (VGG16)
- Validation Dataset
- Binary Crossentropy
- Adam Optimizer
- Early Stopping
- Model Checkpoint
📊 Prediction Pipeline
Upload Image
↓
Read Image
↓
Resize
↓
Preprocess
↓
Load Model
↓
Predict Probability
↓
Threshold = 0.5
↓
NORMAL
or
PNEUMONIA
🌐 Deployment Architecture
User
│
▼
Streamlit Frontend
│
▼
Flask Backend
│
▼
Load Selected Model
│
┌───────────┼────────────┐
▼ ▼ ▼
CNN10 CNN20 VGG16
│ │ │
└───────────┼────────────┘
▼
Prediction Engine
│
▼
Display Result
📂 Project Structure
Chest-XRay-Pneumonia-Detection
│
├── app.py
├── config.py
├── predictor.py
├── utils.py
├── requirements.txt
├── README.md
│
├── models
│ ├── model_xray.h5
│ ├── model_pre.h5
│ └── best_model.keras
│
├── static
│
├── templates
│
├── css
│
├── dataset
│
└── screenshots
▶️ Installation
Clone Repository
git clone https://github.com/yourusername/Chest-XRay-Pneumonia-Detection.git
Install Dependencies
pip install -r requirements.txt
Run Flask
python app.py
Run Streamlit
streamlit run app.py
🖥 Example Prediction
Input
Chest X-Ray Image
↓
Model Prediction
Probability : 0.9821
Prediction :
🦠 PNEUMONIA
📈 Future Improvements
- EfficientNet
- ResNet50
- DenseNet121
- Grad-CAM Heatmaps
- Multi-Class Disease Detection
- DICOM Support
- Cloud Deployment
- Docker
- CI/CD Pipeline
- REST API Authentication
👨💻 Author
Sudheer Muthyala
B.Tech – Electronics and Communication Engineering
Machine Learning | Deep Learning | Computer Vision | Python | Flask | Streamlit
GitHub: https://github.com/M-Sudheer18
⭐ If you found this project helpful
Please consider giving this repository a ⭐ on GitHub. It motivates future improvements and helps others discover the project.
- Downloads last month
- 61