% Set image directory imageDir = 'images/'; % Change this to your image folder path imageFiles = dir(fullfile(imageDir, '*.png')); % Change extension if needed (e.g., .jpg, .bmp) numImages = length(imageFiles);
% Read first image to get dimensions firstImage = imread(fullfile(imageDir, imageFiles(1).name)); if size(firstImage, 3) == 3 firstImage = rgb2gray(firstImage); % Convert to grayscale if RGB end [rows, cols] = size(firstImage); imageSize = rows * cols;
% Create data matrix where each column is an image vector dataMatrix = zeros(imageSize, numImages);
for i = 1:numImages img = imread(fullfile(imageDir, imageFiles(i).name)); if size(img, 3) == 3 img = rgb2gray(img); end img = im2double(img); % Convert to double dataMatrix(:, i) = img(:); % Vectorize image end
% Mean normalization meanImage = mean(dataMatrix, 2); X = dataMatrix - meanImage;
% Compute covariance matrix C = cov(X');
% Eigen decomposition [eigVectors, eigValues] = eig(C); eigValues = diag(eigValues);
% Sort eigenvalues and vectors in descending order [~, idx] = sort(eigValues, 'descend'); eigVectors = eigVectors(:, idx);
% Project data onto principal components k = 10; % Number of principal components to keep eigVectors_k = eigVectors(:, 1:k); projectedData = eigVectors_k' * X;
% Reconstruct an image using top k components reconstructed = eigVectors_k * projectedData meanImage;
% Show original and reconstructed images figure; subplot(1, 2, 1); imshow(reshape(dataMatrix(:, 1), rows, cols), []); title('Original Image');
321 0 0 0
Principle Component Analysis (PCA) Matlab Code (ChatGPT)
Amardip Ghosh #Advanced Propulsion Systems (APSYS) Lab
% PCA on an image set
clc; clear; close all;
% Set image directory
imageDir = 'images/'; % Change this to your image folder path
imageFiles = dir(fullfile(imageDir, '*.png')); % Change extension if needed (e.g., .jpg, .bmp)
numImages = length(imageFiles);
% Read first image to get dimensions
firstImage = imread(fullfile(imageDir, imageFiles(1).name));
if size(firstImage, 3) == 3
firstImage = rgb2gray(firstImage); % Convert to grayscale if RGB
end
[rows, cols] = size(firstImage);
imageSize = rows * cols;
% Create data matrix where each column is an image vector
dataMatrix = zeros(imageSize, numImages);
for i = 1:numImages
img = imread(fullfile(imageDir, imageFiles(i).name));
if size(img, 3) == 3
img = rgb2gray(img);
end
img = im2double(img); % Convert to double
dataMatrix(:, i) = img(:); % Vectorize image
end
% Mean normalization
meanImage = mean(dataMatrix, 2);
X = dataMatrix - meanImage;
% Compute covariance matrix
C = cov(X');
% Eigen decomposition
[eigVectors, eigValues] = eig(C);
eigValues = diag(eigValues);
% Sort eigenvalues and vectors in descending order
[~, idx] = sort(eigValues, 'descend');
eigVectors = eigVectors(:, idx);
% Project data onto principal components
k = 10; % Number of principal components to keep
eigVectors_k = eigVectors(:, 1:k);
projectedData = eigVectors_k' * X;
% Reconstruct an image using top k components
reconstructed = eigVectors_k * projectedData meanImage;
% Show original and reconstructed images
figure;
subplot(1, 2, 1);
imshow(reshape(dataMatrix(:, 1), rows, cols), []);
title('Original Image');
subplot(1, 2, 2);
imshow(reshape(reconstructed(:, 1), rows, cols), []);
title(['Reconstructed with ', num2str(k), ' PCs']);
disp('PCA completed successfully!');