How to extract feature from multiple image?

The aim of the outcome; which to display the result is similar to this:

https://www.mathworks.com/matlabcentral/fileexchange/25157-image-segmentation-tutorial

but instead I would like to fetch the feature from other images, where only one object inside each image.

1. Read image from file - successful

2. Convert RGB to gray - successful

3. Convert Gray to Binary - successful

4. Extract the feature for each image - failed. - It supposed to return Area for each 10 images, somehow it only return 1 value.

[KSSV]

Post your code, so that we can check and find out why you are getting only one return value.

[Farhan Abdul Wahab]

Here's the code:

    close all;
    clear all;
    clc;

    %(1)-- read all images from folder
    path_data_train=strrep(cd,...
    'Classification_Shape','ShapeImage\BottleTraining');

    % data training of Good (jn)
    lots_of_data_train_jn=2;

    % data training of Bad (jl)
    lots_of_data_train_jl=1;

    % initialization of matrix dataset
    n=lots_of_data_train_jn+lots_of_data_train_jl;

    for i=1:n

    if(i<=lots_of_data_train_jn)
        % reading each file of Good
        filename=strcat(path_data_train,'\','Good',...
        num2str(i),'.jpg');
        class{i}='Good';
    else
        % reading each file of Bad
       filename=strcat(path_data_train,'\','Bad',...
        num2str(i-(lots_of_data_train_jn)),'.jpg');
        class{i}='Reject';        
    end

    I = imread (filename);
       if(size(I,3)==4) % resize image
        I(:,:,1)=[]; % convert to I = [MxNx3]
       end
    %figure, imshow (I);

    % (2)-- Convert data to gray
    I_gray=rgb2gray (I);
    %figure,imshow(I_gray); 

    % (3)-- Create inverted binary image
    %I_biner=zeros(size(I_gray,1),size(I_gray,2));
    %I_biner(find(I_gray<255))=1;
    I_biner = im2bw (I_gray, 254/255);
    I_biner = bwareaopen (I_biner, 5); % remove small region less than 5pixels of binary image
    % I_biner = imfill(I_biner, 'holes');
    %figure, imshow(I_biner);  

    % create max filter image from binary image
    %windowing_size must be valued an odd number >=3
    windowing_size=5;
    max_filter_I_biner=Function_MaxFilterBiner_(I_biner,windowing_size); 
    figure, imshow(max_filter_I_biner);

    % draw boundaries line
    hold on;
    boundaries = bwboundaries(max_filter_I_biner);
    numberOfBoundaries = size(boundaries, 1);
    for i = 1 : numberOfBoundaries
        thisBoundary = boundaries{i};
        plot(thisBoundary(:,2), thisBoundary(:,1), 'g', 'LineWidth', 2);
    end
    hold off;

    Measurements = regionprops(max_filter_I_biner, 'Area');
    theArea = [Measurements.Area]

    end

Answer 1

[Image Analyst]

It looks like there are two problem. You are using the loop index "i" inside another, outer loop, which also has the same loop index. Not a wise choice. And it looks like the outer i, which should be renamed to something like imageNumber is the image number but your areas are just for the current image. So put all the areas into a cell. It needs to be a cell because different areas can have different numbers of regions.

theArea{imageNumber} = [Measurements.Area]

By the way, in the MATLAB editor, type control-a control-i to fix up your indenting before you paste it here. It will make your code easier to follow.

[Farhan Abdul Wahab]

Thank you for the looping correction. It works. Now I have the Area quantity equivalent to image quantity. Somehow I'm not clear on your statement of:

" So put all the areas into a cell. It needs to be a cell because different areas can have different numbers of regions."

What does it mean, and how to do it? By using this?

theArea(imageNumber) = [Measurements.Area]

Below is my binary image for Area calculation.

[Image Analyst]

Because the number of areas might vary from image to image, use a cell array and braces:

theArea{imageNumber} = [Measurements.Area]

If you're going to have one blob always, like if you use bwareafilt(binaryImage, 1), then you can use a simple numerical array with parentheses:

theArea(imageNumber) = Measurements.Area;