Pneumonia Classifier

• This pneumonia classifier application takes a chest X-ray image as input and classifies it as either pneumonia or normal.

• The frontend of the application is created with the Python open-source library Streamlit.

• For classification, we use Resnet18, a deep learning model known for image recognition tasks, which has an accuracy of 80% on our test dataset.

• We selected this model after experimenting with conventional machine learning model, custom cnn architecture, vision transformer pretrained model, and various deep learning pretrained models.

• We achieved the highest accuracy of 86% with the Vision Transformer (ViT), but we selected Resnet18 to balance performance and complexity, Resnet18 is generally faster and more lightweight compared to ViT.

• The application is deployed to Hugging Face Spaces and is now live. You can access it here: Pneumonia Classifier

Dataset Overview

• The dataset used are Chest X-Ray Images (Pneumonia).
• It have total 5,863 X-Ray images (JPEG).
• The dataset is organized into three main folders: train, val and test.
• The dataset consist of two categories:
  1. PNEUMONIA: 3875 images
  2. NORMAL: 1341 images

---

Models Accuracy Table

Model	Accuracy (%)
Conventional Machine Learning (SVM)	0.75
Custom CNN	0.80
VGG16	0.79
EfficientNet-B0	0.80
Resnet18	0.80
ViT	0.86

Models Accuracy Comparison

---

Models Training Time Comparison

---

Confusion Matrices of Models

🚀 Quick Start

1. Clone the Repository:

   git clone https://github.com/ImranNawar/pneumonia-classifier.git

2. Navigate to the Directory:

   cd pneumonia-classifier

3. Install Dependencies:

   pip install requirements.txt

4. Train the Models:

   • Open the notebook pneumonia-classifer.ipynb
   • Run all the cells to train all the models or just specific cells, if you want to train or finetune a specific model.

5. Run the Application:

   cd src
   streamlit run main.py

📁 Project Structure

pneumonia-classifier/
├── content/
│   ├── plots/
│   ├── app_interface.png
│   └── background.jpg
├── data/                               # ignored in .gitignore
│   ├── test/
│   │   ├── NORMAL/
│   │   └── PNEUMONIA/
│   ├── train/
│   │   ├── NORMAL/
│   │   └── PNEUMONIA/
│   └── val/
│       ├── NORMAL/
│       └── PNEUMONIA/
├── model/
│   ├── label.txt
│   └── resnet18.pth                     # ignored in .gitignore
├── src/
│   ├── main.py
│   └── util.py
├── training/
│   └── pneumonia-classifier.ipynb
├── .gitignore
├── LICENSE
├── README.md
└── requirements.txt

📄 License

This project is licensed under the MIT License.