Pneumonia Detection using Chest X-ray Images

This project uses a Convolutional Neural Network (CNN) to classify chest X-ray images as either pneumonia-positive or normal. The model is trained and evaluated on the Chest X-ray Images (Pneumonia) dataset by Paul Timothy Mooney, available on Kaggle.

Dataset

The dataset contains X-ray images divided into train, validation, and test sets. Images are preprocessed and augmented to improve model generalization.

Directory Structure

• train/: Contains training images.
• val/: Contains validation images.
• test/: Contains test images.

Ensure these directories match the paths specified in the script.

Installation

1. Clone this repository and navigate to the project directory.
2. Install the required libraries:

   pip install -r requirements.txt

3.Download the dataset from Kaggle using

kaggle datasets download -d paultimothymooney/chest-xray-pneumonia

Project Structure

• Data Augmentation: The project uses ImageDataGenerator from Keras to apply transformations such as rotations, zooms, and flips to increase dataset diversity.
• Model Architecture: A CNN model with binary classification capabilities is built and compiled using Keras. Training is monitored using TensorBoard.
• Callbacks: Early stopping and learning rate adjustments are implemented to optimize training performance.

Results

Saved model weights are stored in best_model.keras. Training and validation metrics are plotted for loss and accuracy.

Acknowledgments

Dataset: Kaggle - Chest X-ray Images (Pneumonia)