MATLAB: How can i extract and save LBP features of multiple images, so that i can use them as a DB features

cbirImage Processing Toolboxlbpfeatureslocal binary pattern

Hii, I have a folder of 200 images, I want to extract the LBP features of each image and save them in an Excel sheet so that I can use them as a DB features and retrieve, based on them, the accurate images in a CBIR project. I used this code for the extraction but the output gives me the feature vector of only one image.

image_folder = 'C:\Users\Desktop\images\resized';
ext_img = '*.png';              
files = dir(fullfile(image_folder, ext_img)); 
assert(numel(files) > 0, 'No file was found. Check that the path is correct');
my_img = struct('img', cell(size(files)));  
for i = 1:numel(files) 
Img = imread(fullfile(image_folder, files(i).name));
gray = rgb2gray(Img);
LBP_image = extractLBPFeatures(gray,'Cellsize',[32 32], 'Normaas i lization','None');
numNeigbors = 8 %%Reshape the LBP features into a number of neighbors -by- number of cells array to access histograms for each individual cell.
numBins = numNeigbors*(numNeigbors-1)+3;
Cell_Hists = reshape(LBP_image,numBins,[]);
Cell_Hists = bsxfun(@rdivide,Cell_Hists,sum(Cell_Hists));%Normalize each LBP cell histogram using L1 norm
LBP_image = reshape(Cell_Hists,1,[]);%Reshape the LBP features vector back to 1-by- N feature vector
end

My second question is how to store them in Excel (in order to have the feature vectors of the 200 images). I used the xlswrite() function but as I said it gives me the vector of only one images (the last one) with a size of 1×3776.

Best Answer

That code is not very robust at all. Mainly you need to preallocate a matrix for LPB_image and load the results for each image into the proper row of LBP_image. And only write the workbook if there were some files found. So you need to do (as a start):

image_folder = 'C:\Users\Desktop\images\resized';
if ~isfolder(image_folder)
	errorMessage = sprintf('Error: Folder not found:\n%s', image_folder);
	fprintf('%s\n', errorMessage);
	errordlg(errorMessage);
	return;
end
ext_img = '*.png';
files = dir(fullfile(image_folder, ext_img));
assert(numel(files) > 0, 'No file was found. Check that the path is correct');
my_img = struct('img', cell(size(files)));
numberOfFiles = numel(files);
LBP_image = zeros(numberOfFiles, 3776); % Preallocate a row for every file.
for k = 1 : numberOfFiles
	baseFileName = files(k).name;
	fullFileName = fullfile(files(k).folder, baseFileName);
	fprintf('Processing image #%d of %d : "%s"\n', k, numberOfFiles, baseFileName);
	thisImage = imread(fullFileName);
	% Convert to gray scale, if needed.
	if ndims(thisImage) > 2
		grayImage = rgb2gray(thisImage);
	else
		% Already gray scale so no conversion needed.
		grayImage = thisImage;
	end
	% Compute LBP features.
	LBP_image = extractLBPFeatures(grayImage, 'Cellsize',[32 32], 'Normaas i lization','None');
	numNeigbors = 8; %%Reshape the LBP features into a number of neighbors -by- number of cells array to access histograms for each individual cell.
	numBins = numNeigbors * (numNeigbors-1) + 3;
	Cell_Hists = reshape(LBP_image,numBins,[]);
	Cell_Hists = bsxfun(@rdivide,Cell_Hists,sum(Cell_Hists));%Normalize each LBP cell histogram using L1 norm
	% Assign the results of this image, a 3776 row vector, to the proper row in our output.
	LBP_image(k, :) = reshape(Cell_Hists,1,[]);%Reshape the LBP features vector back to 1-by- N feature vector
end
if numberOfFiles >= 0
	xlswrite(fileName, LBP_image);
end

Of course there is lots more you could do to improve this code (add comments, alert the user if no files were found, etc.)

Related Solutions

MATLAB: What is the mistale in the loop

Try this:

clc;    % Clear the command window.
close all;  % Close all figures (except those of imtool.)
clear;  % Erase all existing variables. Or clearvars if you want.
workspace;  % Make sure the workspace panel is showing.
format long g;
format compact;
fontSize = 25;
% Read in a standard MATLAB gray scale demo image.
folder = fullfile(matlabroot, '\toolbox\images\imdemos');
baseFileName = 'cameraman.tif';
% Get the full filename, with path prepended.
fullFileName = fullfile(folder, baseFileName);
% Check if file exists.
if ~exist(fullFileName, 'file')
  % File doesn't exist -- didn't find it there.  Check the search path for it.
  fullFileNameOnSearchPath = baseFileName; % No path this time.
  if ~exist(fullFileNameOnSearchPath, 'file')
    % Still didn't find it.  Alert user.
    errorMessage = sprintf('Error: %s does not exist in the search path folders.', fullFileName);
    uiwait(warndlg(errorMessage));
    return;
  end
end
referenceGrayImage = imread(fullFileName);
% Get the dimensions of the image.
% numberOfColorBands should be = 1.
[rows, columns, numberOfColorBands] = size(referenceGrayImage);
if numberOfColorBands > 1
  % It's not really gray scale like we expected - it's color.
  % Convert it to gray scale by taking only the green channel.
  referenceGrayImage = referenceGrayImage(:, :, 2); % Take green channel.
end
% Display the original gray scale image.

subplot(2, 2, 1);
imshow(referenceGrayImage, []);
title('Original Grayscale Image', 'FontSize', fontSize);
% Enlarge figure to full screen.
set(gcf, 'Units', 'Normalized', 'OuterPosition', [0 0 1 1]);
% Give a name to the title bar.
set(gcf, 'Name', 'Demo by ImageAnalyst', 'NumberTitle', 'Off')
% Let's compute and display the histogram.
[pixelCount, grayLevels] = imhist(referenceGrayImage);
subplot(2, 2, 2);
bar(grayLevels, pixelCount);
grid on;
title('Histogram of original image', 'FontSize', fontSize);
xlim([0 grayLevels(end)]); % Scale x axis manually.
srcFiles = dir(fullfile(folder, '*.jpg')); %directory of the source folder
Im1 = im2double(referenceGrayImage); % Calculate the Normalized Histogram
hn1 = imhist(referenceGrayImage)./numel(referenceGrayImage);
for i = 1 : length(srcFiles)      % determins how many images are there in the folder,
  filename = fullfile(folder,srcFiles(i).name)
  thisImage = imread(filename);
  % Display the original gray scale image.
  subplot(2, 2, 3);
  imshow(thisImage);
  title(filename, 'FontSize', fontSize);
  drawnow;
  Im2=rgb2gray(thisImage);
  Im2 = im2double(Im2);
  hn2 = imhist(Im2)./numel(Im2);
  
  % Calculate the histogram error

  f(i) = sum((hn1 - hn2).^2);% Calculate the histogram error
  f(i) %display the result to console  
  
  promptMessage = sprintf('f=%f\nDo you want to Continue processing,\nor Cancel to abort processing?',f(i));
  titleBarCaption = 'Continue?';
  button = questdlg(promptMessage, titleBarCaption, 'Continue', 'Cancel', 'Continue');
  if strcmpi(button, 'Cancel')
    break;
  end
end

MATLAB: Is it possible to retrieve similar images by using this code

No, there is a lot more to CBIR than similarity of histograms. Imagine you had an image of a bird and you took the pixels and rearranged them. Let's say that you sorted them and rearranged them into a ramp or gradient from one color to another. It would look like a smooth image - nothing at all like the original image of a bird - yet it would have the very same histogram of the bird image. Your code would return that image as being similar to a bird. That is probably not what you'd want. CBIR is a complicated and rapdily evolving subject that tags each image in the database with lots of information. You should do a search on CBIR to learn about this subject. You'll see that there's no way it can be done with a short, 100-line long MATLAB script.

Best Answer

Related Solutions

MATLAB: What is the mistale in the loop

MATLAB: Is it possible to retrieve similar images by using this code

Related Question