Eigenfaces

This repository contains code and resources for a mini-project in the Statistical signal and data processing through applications course.

Features

• Image Preprocessing: Utilities for reshaping, normalizing, and preparing image datasets.
• Feature Extraction:
  • PCA (Principal Component Analysis) for dimensionality reduction
  • Phase congruency and other custom feature maps
• Classification Pipelines:
  • Logistic Regression
  • SVM (Support Vector Machine)
  • Linear Regression (for regression tasks)
  • Support for custom classifiers
• Evaluation & Visualization:
  • Confusion matrix, F1 score, and accuracy metrics
  • Visualization of PCA eigenvalues and feature maps
  • Customizable plotting with adjustable font sizes

Getting Started

Prerequisites

• Python 3.8+
• PyTorch
• scikit-learn
• matplotlib
• facenet-pytorch
• Pillow
• numpy

Install dependencies with:

pip install torch torchvision scikit-learn matplotlib facenet-pytorch pillow numpy

Example Usage

PCA Feature Extraction:

from sklearn.decomposition import PCA

pca = PCA(n_components=100, whiten=True)
X_train_pca = pca.fit_transform(X_train)
X_test_pca = pca.transform(X_test)

Classification:

from sklearn.linear_model import LogisticRegression

clf = LogisticRegression(max_iter=1000)
clf.fit(X_train_pca, y_train)
y_pred = clf.predict(X_test_pca)

Folder Structure

.
├── images.mat/                # Yale Face Database images
├── classes.mat/                # Yale image labels
├── generated_images.mat/                # Synthetic dataset images
├── generated_classes.mat/                # Synthetic dataset labels
├── Generate_data.ipynb/                # A script for generating synthetic dataset of random images
├── generate_results.ipynb/                # A script to generate classification results for the report
├── PCA.ipynb/                # Implementation of the PCA based classification
├── Fisherfaces.ipynb/                # Implementation of the Fisherfaces based classification
├── Gabor-Fisher.ipynb/                # Implementation of the Gabor filter based classification
├── README.md