Cat Image Classification Neural Network

This project implements a neural network using numpy and scipy to classify cat images from non-cat images. The neural network is built and trained in a Jupyter Notebook. The dataset used for training and testing consists of cat and non-cat images.

Overview

The neural network architecture comprises several layers, including linear, ReLU (Rectified Linear Unit), and sigmoid activation functions. It is implemented as a 4-layer neural network with customizable layer dimensions.

Project Structure

• cat_image_classification.ipynb: Jupyter Notebook containing the code for the neural network implementation, training, and evaluation.
• datasets/: Directory containing the dataset for training and testing (train_catvnoncat.h5 and test_catvnoncat.h5).

Dependencies

The project relies on the following libraries:

• NumPy
• SciPy
• h5py
• Matplotlib
• PIL (Python Imaging Library)

Usage

1. Dataset Preparation: Ensure the dataset (train_catvnoncat.h5 and test_catvnoncat.h5) is placed in the datasets/ directory.
2. Jupyter Notebook Execution: Open and execute the cat_image_classification.ipynb Jupyter Notebook to train and evaluate the neural network.
3. Modify Parameters (Optional): Modify the neural network architecture, learning rate, or number of iterations in the notebook for experimentation.

Results

The trained neural network achieves a certain level of accuracy in classifying cat images from non-cat images. The notebook provides visualizations of the training process, accuracy, and misclassified images.

Rendered Notebook Site

The rendered site of the notebook can be accessed here.

Contributing

Contributions, issues, and feature requests are welcome. Feel free to submit a pull request or open an issue for any suggestions or improvements.

License

This project is licensed under the MIT License.