Lung Disease Classification from Chest X-Ray Images This project aims to train an AI model to recognize lung diseases from chest X-ray images using the NIH Chest X-ray dataset. The model leverages deep learning techniques and transfer learning to perform multi-label classification of 14 different lung diseases.

Table of Contents Overview Dataset Prerequisites Installation Dataset Download Option 1: Download Individual Files via Kaggle API Option 2: Download via Kaggle Web Interface Option 3: Use a Smaller Sample Dataset Project Structure Running the Code Troubleshooting Additional Resources Acknowledgements Overview This project involves:

Setting up a development environment with necessary libraries. Downloading and preprocessing the NIH Chest X-ray dataset. Building a convolutional neural network (CNN) model using transfer learning (DenseNet121). Training the model for multi-label classification. Evaluating the model's performance. Saving the trained model for future use. Dataset NIH Chest X-ray Dataset Source: NIH Chest X-ray Dataset Description: Contains over 112,000 chest X-ray images with labels for 14 different diseases and one "No findings" class. Data Limitations: Labels were extracted using NLP from radiology reports (estimated accuracy >90%). Some labels may be erroneous due to NLP extraction errors. Very limited numbers of disease region bounding boxes. Classes: Atelectasis Consolidation Infiltration Pneumothorax Edema Emphysema Fibrosis Effusion Pneumonia Pleural Thickening Cardiomegaly Nodule Mass Hernia No Finding Prerequisites Python: 3.6 or higher pip: Package installer for Python Git: For version control (optional but recommended) Installation

1. Clone the Repository bash Copy code git clone https://github.com/your-username/your-repo-name.git cd your-repo-name
2. Set Up a Virtual Environment Using venv:

bash Copy code python -m venv venv Activate the Virtual Environment:

Windows:

bash Copy code venv\Scripts\activate macOS/Linux:

bash Copy code source venv/bin/activate 3. Upgrade pip bash Copy code pip install --upgrade pip 4. Install Required Libraries bash Copy code pip install -r requirements.txt Contents of requirements.txt:

Copy code numpy pandas tensorflow scikit-learn matplotlib seaborn Dataset Download Due to the large size of the NIH Chest X-ray dataset (~42 GB), downloading it requires careful handling. Below are options to download the dataset:

Option 1: Download Individual Files via Kaggle API Step 1: Install Kaggle API bash Copy code pip install kaggle Step 2: Set Up Kaggle API Credentials Obtain Kaggle API Token:

Log in to your Kaggle account. Go to Account Settings. Click on Create New API Token. A kaggle.json file will be downloaded. Place kaggle.json in the Correct Directory:

Windows:

bash Copy code mkdir %USERPROFILE%.kaggle move kaggle.json %USERPROFILE%.kaggle macOS/Linux:

bash Copy code mkdir ~/.kaggle mv kaggle.json ~/.kaggle/ chmod 600 ~/.kaggle/kaggle.json Step 3: Accept the Dataset's Terms Navigate to the NIH Chest X-ray Dataset. Click Download and accept any terms and conditions if prompted. Step 4: Download Individual Files List Available Files:

bash Copy code kaggle datasets files nih-chest-xrays/data Download Necessary Files:

bash Copy code

Create dataset directory if it doesn't exist

mkdir dataset

Download Data_Entry_2017.csv

kaggle datasets download -d nih-chest-xrays/data -f Data_Entry_2017.csv -p dataset/

Download image zip files (adjust the range if you want fewer files)

for i in $(seq -f "%03g" 1 12); do kaggle datasets download -d nih-chest-xrays/data -f images_${i}.zip -p dataset/ done Step 5: Extract the Zip Files bash Copy code cd dataset/ mkdir images

Extract each images_*.zip file into the images/ directory

for file in images_*.zip; do unzip "$file" -d images/ done Option 2: Download via Kaggle Web Interface Navigate to the NIH Chest X-ray Dataset.

Scroll down to the Data Explorer section.

Download individual files (e.g., Data_Entry_2017.csv, images_001.zip, etc.) by clicking on them and then clicking the Download button.

Move the downloaded files into the dataset/ directory.

Extract the zip files as described in Step 5 above.

Option 3: Use a Smaller Sample Dataset If you prefer working with a smaller dataset initially:

Download the Sample Dataset bash Copy code kaggle datasets download -d nih-chest-xrays/sample -p dataset/ Extract the Sample Dataset bash Copy code cd dataset/ unzip sample.zip -d images/ Adjust Code for Sample Dataset Update paths in main.py to point to the sample dataset. Use sample_labels.csv provided in the sample dataset. Project Structure scss Copy code project_directory/ ├── dataset/ │ ├── Data_Entry_2017.csv │ ├── images/ │ │ ├── image1.png │ │ ├── image2.png │ │ └── ... │ └── (other dataset files) ├── main.py ├── requirements.txt └── venv/ ├── bin/ ├── lib/ └── (virtual environment files) Running the Code

1. Activate the Virtual Environment Windows:

bash Copy code venv\Scripts\activate macOS/Linux:

bash Copy code source venv/bin/activate 2. Run the Script bash Copy code python main.py 3. Monitor Training Training progress, including loss and accuracy, will be displayed in the console. After training, evaluation metrics and confusion matrices will be displayed.