How to find PIQE and BRISQUE for 500 images and how to segregate that with the threshold of min, max and median???

1. how to find BRISQUE (Blind/Referenceless Image Spatial Quality Evaluator) and PIQE (Perception based Image Quality Evaluator) for 500 images in a folder.
2. how to segregate these images with threshold of minimum, maximum and median values in different folders.

(eg.,

Answer 1

[Vaibhav]

Hi Vimal,

It is my understanding that clarification is sought concerning the calculation of “PIQE” and “BRISQUE” for 500 images, and subsequently organizing them into various folders.

The below code offers a solution to the question of finding “BRISQUE” and “PIQE” scores for a set of images in a folder. It calculates the quality scores for 100 randomly selected images from the “CIFAR-10” dataset and segregates them into different folders based on the specified thresholds. The “CIFAR-10” dataset used in the code can be replaced by another dataset according to the specific requirements. By adapting the code and replacing the “CIFAR-10” dataset with concerned image dataset, evaluation of the quality of your images and organizing them based on their quality levels can be achieved.

Please find the code for reference below:

% Download and extract the “CIFAR-10” dataset
url = 'https://www.cs.toronto.edu/~kriz/cifar-10-matlab.tar.gz';
downloadFolder = fullfile(tempdir, 'cifar10');
filename = fullfile(downloadFolder, 'cifar-10-matlab.tar.gz');
untar(url, downloadFolder);

% Load the “CIFAR-10” dataset
load(fullfile(downloadFolder, 'cifar-10-batches-mat', 'test_batch.mat'));

% Select images randomly from the dataset
numImages = 100; % Change to 500 if desired
selectedIndices = randperm(size(data, 1), numImages);

% Initialize arrays to store the BRISQUE and PIQE scores
brisqueScores = zeros(numImages, 1);
piqeScores = zeros(numImages, 1);

% Loop through each selected image and calculate the “BRISQUE” and “PIQE” scores
for i = 1:numImages
    img = reshape(data(selectedIndices(i), :), 32, 32, 3);
    
    % Calculate BRISQUE score
    brisqueScores(i) = brisque(img);
    
    % Calculate “PIQE” score
    piqeScores(i) = piqe(img);
end

% Calculate the minimum, maximum, and median values of the “BRISQUE” and “PIQE” scores
minBrisque = min(brisqueScores);
maxBrisque = max(brisqueScores);
medianBrisque = median(brisqueScores);

minPiqe = min(piqeScores);
maxPiqe = max(piqeScores);
medianPiqe = median(piqeScores);

% Create separate folders for images based on the threshold values
outputFolder = pwd; % Use the present working directory
mkdir(outputFolder, 'BelowMinBrisque');
mkdir(outputFolder, 'AboveMaxBrisque');
mkdir(outputFolder, 'BetweenMinMaxBrisque');
mkdir(outputFolder, 'BelowMinPiqe');
mkdir(outputFolder, 'AboveMaxPiqe');
mkdir(outputFolder, 'BetweenMinMaxPiqe');

% Move the images to the appropriate folders based on the threshold values
for i = 1:numImages
    img = reshape(data(selectedIndices(i), :), 32, 32, 3);
    imgName = sprintf('image%d.jpg', i);
    
    % Move images based on BRISQUE threshold
    if brisqueScores(i) < minBrisque
        imwrite(img, fullfile(outputFolder, 'BelowMinBrisque', imgName));
    elseif brisqueScores(i) > maxBrisque
        imwrite(img, fullfile(outputFolder, 'AboveMaxBrisque', imgName));
    else
        imwrite(img, fullfile(outputFolder, 'BetweenMinMaxBrisque', imgName));
    end
    
    % Move images based on PIQE threshold
    if piqeScores(i) < minPiqe
        imwrite(img, fullfile(outputFolder, 'BelowMinPiqe', imgName));
    elseif piqeScores(i) > maxPiqe
        imwrite(img, fullfile(outputFolder, 'AboveMaxPiqe', imgName));
    else
        imwrite(img, fullfile(outputFolder, 'BetweenMinMaxPiqe', imgName));
    end
end

Please refer to the following documentations for more information:

https://in.mathworks.com/help/images/ref/brisque.html

https://in.mathworks.com/help/images/ref/piqe.html

Hope this helps!

Regards,

Vaibhav

[Vimal Raj]

Thank You Sir.