Neural Network from Scratch for Pneumonia Detection

This repository contains a simple implementation of a neural network built from scratch using Numpy. The neural network is trained to classify pneumonia images into two categories: normal and pneumonia.

Usage

Clone the repository:

git clone https://github.com/tu_usuario/tu_repositorio.git

Navigate to the project directory:
```
cd tu_repositorio
```
Install the required dependencies (it is recommended to use a virtual environment):
```
pip install -r requirements.txt
```
Run the main script to train the neural network:
```
python main.py
```

Dataset

The dataset consists of X-ray images of lungs with and without pneumonia. The images are categorized into two classes: normal and pneumonia.

You can access the dataset on Kaggle.

Structure

The project is organized as follows:

neural_network: Contains the implementation of the neural network.
- neural_network.py: Defines the NeuralNetwork class with methods for training and inference.
- create_data.py: Provides functionality for loading and formatting the dataset.
main.py: Script to train the neural network using the dataset.

Dependencies

Python 3
Numpy
OpenCV
Matplotlib
Scipy (Only for the sigmoid equation) TODO: Do the sigmoid equation only with Numpy

License

This project is licensed under the MIT License - see the LICENSE file for details.

Author Santiago (santiagosaav.99@gmail.com)