can I store images with differing sizes into a matrix in matrix?

I'm doing object detection and classification to either human or non human from a surveillance video clip. when I detect object and plot it in a rectangular box,I need to save the cropped portion of this object for the purpose of feature extraction and further training the classifier.The problem I encounter is, I'm not able to save the cropped images to a matrix;as it is not copied,I could get only the last cropped image.But I need all the blobs detected. Can anyone please help me to come out of this? Please.... Each cropped image is of different size....

Answer 1

[Jan]

It would be possible to find the problem is your code, when you post the code.

A general method to store arrays of different sizes:

n = 10;
Data = cell(1, n);
for k = 1:n
  Data{k} = rand(n, n);
end

[soumya]

source = VideoReader('Z:\\M.Tech\\s3\\NIC\\mfiles\\camclips\\tes.avi'); z=1; cropImgs=cell(:,:,z); for n=1:source.NumberOfFrames bgImg=sprintf('Z:\\minipro\\mog\\pic%d.jpg',n); bgeliminatedImg=imread(bgImg); originalImg=read(source,n); if true % code end % Method for thresholding: using a logical operation. thresholdValue = 100; binaryImage = bgeliminatedImg > thresholdValue; % Bright objects will be the chosen if you use >

    % Do a "hole fill" to get rid of any background pixels inside the blobs.
    binaryImage = imfill(binaryImage, 'holes');

    I=binaryImage;

    L = bwlabel(I, 8);
    RGB= label2rgb (L, 'hsv', 'k', 'shuffle'); % pseudo random color labels
    subplot(2, 2, 1); imagesc(L); title('Labeled Image,bwlabel()'); axis square;hold on;

    % Maximize the figure window.
    set(gcf, 'Position', get(0, 'ScreenSize'));
    subplot(2, 2, 2); imagesc(RGB); title('Conn-components-colored labels,label2rgb()'); axis square;hold on;

    blobMeasurements = regionprops(L,bgeliminatedImg, 'all');     
    numberOfBlobs = size(blobMeasurements, 1);
    subplot(2, 2, 3); imagesc(I); title('Outlines, from bwboundaries()'); axis square;
    hold on;

    boundaries = bwboundaries(I,'noholes');

    for k = 1 : numberOfBlobs
       % if(blobMeasurements(k).Area>4000)  %%to avoid very small connected components 
            thisBoundary = boundaries{k};
            plot(thisBoundary(:,2), thisBoundary(:,1), 'g', 'LineWidth', 2);                      
       % end

    end

    allBlobAreas = [blobMeasurements.Area];
    largeBlobs = find(allBlobAreas > 5000);  % Take the larger objects.
    % we use ismember to select the blobs that meet our criteria.
    newImg = ismember(L, largeBlobs); 
    maskedImg = bgeliminatedImg; % Simply a copy at first.
    maskedImg(~newImg) = 0;% Set all unwanted pixels to zero.
    %maskedImg(newImg)=1;
     maskedImg=imfill(maskedImg,'holes');
     maskedImg=im2bw(maskedImg,0.5);

    subplot(2, 2, 4);imagesc(maskedImg);hold on;

    title('Original with bounding boxes');
    for k = 1 : numberOfBlobs % Loop through all blobs.
    % Find the mean of each blob. 
        if(largeBlobs==k) 
                thisBlobsBox = blobMeasurements(k).BoundingBox; % Get list ofpixels in current blob.                
                rectangle('Position',thisBlobsBox,'EdgeColor','g');axis square;
                CropImgs{z}=imcrop(originalImg,thisBlobsBox); 
                z=z+1;
        end
    end
    MaskObj(n)=getframe(gcf); 
end

[Jan]

The code is nicer, when you format it. Which part of the code concerns your problem? Which variable does not contain the wanted values?

[soumya]

hey....it works...actually,i dnt used CropImgs{z}=imcrop() statement in my code.instead(CropImgs=imcrop())I included it for u...when you replied,I just thought of trying that.... Thanks a lot sir...... Thank you so much.....