Disease-Diagnosis-Using-AI-Machine-Learning-with-PyTorch Public

This article demonstrates disease diagnosis using AI and machine learning (ML) techniques, specifically leveraging Convolutional Neural Networks (CNNs) with PyTorch. The model classifies medical images, such as chest X-rays, into categories like NORMAL and PNEUMONIA. It highlights the application of AI in healthcare for disease detection.

In this code, we classify whether a patient has pneumonia or not based on chest X-ray images. However, the same approach can be extended to multi-class classification tasks, as long as your input dataset contains labeled images.

For our case, the dataset contains chest X-ray images categorized into two folders: NORMAL and PNEUMONIA. The dataset structure looks like this:

dataset/
├── NORMAL/
│   ├── img1.jpg
│   ├── img2.jpg
│   └── ...
├── PNEUMONIA/
│   ├── img1.jpg
│   └── ...

If you have more classes, simply add more folders under the dataset/ directory, each representing a separate class with its own set of images.

In the code, images are converted into high-dimensional tensors using PyTorch. A Convolutional Neural Network (CNN) — such as ResNet — is used to extract features and perform classification.

The images are normalized to have zero mean and unit variance, which improves training stability and model performance.

The model uses:

• CrossEntropyLoss — a standard loss function for multi-class classification that compares predicted class probabilities (via softmax) with the true labels.
• Adam optimizer — an efficient optimization algorithm that adapts learning rates for each parameter during training.

Code

Visit Github

Further Reading

1.