Dimensionality Reduction with PCA for Image Reconstruction

Project Overview

This project demonstrates the application of Principal Component Analysis (PCA) for dimensionality reduction and image reconstruction. PCA is a powerful statistical technique used to reduce the number of features in a dataset while preserving as much variance (information) as possible. In this project, PCA is applied to a set of images to reduce their dimensionality and reconstruct them to visualize how well the technique retains information.

Components

1. Preprocessing (preprocess.py):

   • Loads images from a specified folder.
   • Resizes and flattens images to prepare them for PCA.

2. PCA Application (pca.py):

   • Applies PCA to the image data to reduce dimensionality.
   • Transforms the original images into a lower-dimensional space and allows for their reconstruction.

3. Visualization (visualize.py):

   • Reconstructs images from the reduced dimensions.
   • Visualizes and compares the original and reconstructed images to evaluate PCA performance.

4. Main Execution (main.py):

   • Integrates preprocessing, PCA application, and visualization.
   • Prints the explained variance ratio to indicate how much information is preserved.

Requirements

• Python 3.x
• numpy
• scikit-learn
• pillow
• matplotlib

Install the required libraries using:

pip install numpy scikit-learn pillow matplotlib

Usage Instructions

Prepare Images

1. Place your images in a folder.
2. Ensure the images are in a format supported by Pillow (e.g., PNG, JPEG).

Update main.py

Modify the image_folder variable to point to the folder containing your images.

image_folder = 'path_to_your_image_folder'

Run The Project

1. Execute the script with:

python3 main.py

View Results

1. The console will display:

• Number of samples
• Number of features
• Explained variance ratio

Key Notes

Dimensionality and Components

• The n_components parameter in PCA determines the number of principal components to use. Adjust this parameter to balance between dimensionality reduction and reconstruction quality.
• Too few components may lead to blurred reconstructions.
• Too many components may not effectively reduce dimensionality.

Image Quality

• Ensure that you have a sufficient number of images for meaningful PCA analysis. The quality of reconstructed images depends on the number of principal components used.

Troubleshooting

• If the reconstructed images are blurred or not clear, consider increasing the number of components used in PCA.

Contribution

• If you would like to contribute to this project, please fork the repository and submit a pull request.