How Can I do image extraction of chosen region?

I have an mammogram image . I use roipoly command to draw tumor suspicious area . Maybe something easier ways to draw that blue polygon.

My question is I need to extract that blue polygon tumor suspicious region from original image and then I will segment that extracted image.

How Can I do that ?

(The image was taken from mias database).

Answer 1

[Image Analyst]

Once you have the mask, do this:

% Mask the image using bsxfun() function to multiply the mask by each channel individually.  
% Works for gray scale as well as RGB Color images.
maskedImage = bsxfun(@times, rgbImage, cast(mask, 'like', rgbImage));

changing the names to whatever you want to call them.

[Image Analyst]

If you also want to crop it out to a smaller image, you can do

props = regionprops(mask, 'BoundingBox'); % Get bounding box of the polygon.
croppedImage = imcrop(maskedImage, props.BoundingBox); % Crop it out.

[Ali Zulfikaroglu]

thank you Image analyst the first bsxfun function worked really well.

But second regionprops function I added but I couldn't see anything. Maybe I couldn't run it.

My matlab command like that .

I also added image.

clear all;
clc;

img=imread('image.jpg');
if(size(img,3)>1)
img=rgb2gray(img);
end

binary_mask=roipoly(img);
figure, imshow(binary_mask);
% I added in here your functions

% Mask the image using bsxfun() function to multiply the mask by each channel individually.  
% Works for gray scale as well as RGB Color images.
maskedImage = bsxfun(@times, img, cast(binary_mask, 'like', img));
figure, imshow(maskedImage);
%Till now , it works. I got what I need 
% but when I add bottow commands didn't see what happen.
 
props = regionprops(binary_mask, 'BoundingBox'); % Get bounding box of the polygon.
croppedImage = imcrop(maskedImage, props.BoundingBox); % Crop it out.

[Image Analyst]

regionprops() doesn't show anything unless you leave off the semicolon. Then to see the images you need to call imshow()

imshow(mask); % Show mask (briefly)
drawnow;
props = regionprops(mask, 'BoundingBox') % Get bounding box of the polygon.
croppedImage = imcrop(maskedImage, props.BoundingBox); % Crop it out.
imshow(croppedImage, []); % Show the cropped image.
drawnow;

[Ali Zulfikaroglu]

Thank you Image Analyst, it helped a lot .

[Ali Zulfikaroglu]

My last question about this subject. You saw I draw blue polygon for tumour suspicious area. But what if more than 1 tumour suspicious area happen , How will draw extra polygon ? Because with roipoly command , only 1 polygon I can draw , it doesn't allow second or third for same image at the same time. Anything else command for annotation/labeling that region for when there is a multiple tumour?

[Image Analyst]

Put it in a loop where you ask the use if they want to draw a polygon and then OR it with the master one

mask = false(rows, columns);
drawAnother = true;
while drawAnother
    thisMask = roipolyold();
    mask = mask | thisMask;
    % ask user if they want to draw another with questdlg().
    promptMessage = sprintf('Do you want to Continue processing,\nor Quit processing?');
    titleBarCaption = 'Continue?';
    buttonText = questdlg(promptMessage, titleBarCaption, 'Continue', 'Quit', 'Continue');
    if contains(buttonText, 'Quit', 'IgnoreCase', true)
    	drawAnother = false;
    end
end

[Ali Zulfikaroglu]

clear all;
clc;

img=imread('image.jpg');
if(size(img,3)>1)
img=rgb2gray(img);
end

mask=roipoly(img);
figure, imshow(mask);
% % Mask the image using bsxfun() function to multiply the mask by each channel individually.  
% % Works for gray scale as well as RGB Color images.
  maskedImage = bsxfun(@times, img, cast(mask, 'like', img));
  figure, imshow(maskedImage);
 
  mask = false(rows, columns);
drawAnother = true;
while drawAnother
    thisMask = roipolyold();
    mask = mask | thisMask;
    % ask user if they want to draw another with questdlg().
    promptMessage = sprintf('Do you want to Continue processing,\nor Quit processing?');
    titleBarCaption = 'Continue?';
    buttonText = questdlg(promptMessage, titleBarCaption, 'Continue', 'Quit', 'Continue');
    if contains(buttonText, 'Quit', 'IgnoreCase', true)
    	drawAnother = false;
    end
end

% I made codes like that and get that fault below

[Ali Zulfikaroglu]

'rows' requires Database Toolbox.

Error in regionn_growing (line 16)

mask = false(rows, columns);

[Image Analyst]

"rows" is the number of rows in your image. You can use size() to get it:

[rows, columns, numberOfColorChannels] = size(img)

[Ali Zulfikaroglu]

I tried and it gave this message

"Warning: Function ROIPOLYOLD will be removed in the future. Use ROIPOLY instead. "

And then I used roipoly function instead of roipolyold. But,

I can't draw second polygon on my first image. I draw polygon on first image (img) and then masked it and then extracted masked image from original image and then it let me draw polygon on extracted image . But my aim is I wanna draw second, third or multi blue polygon on original image (img) at the same time and then extract them together. How will I do it?

[Image Analyst]

I already told you:

mask = false(rows, columns);
drawAnother = true;
while drawAnother
    thisMask = roipoly(); % or however you want to get it.
    mask = mask | thisMask;
    % ask user if they want to draw another with questdlg().
    promptMessage = sprintf('Do you want to Continue processing,\nor Quit processing?');
    titleBarCaption = 'Continue?';
    buttonText = questdlg(promptMessage, titleBarCaption, 'Continue', 'Quit', 'Continue');
    if contains(buttonText, 'Quit', 'IgnoreCase', true)
    	drawAnother = false;
    end
end

Now mask will have several polygon regions on it and you can just pass it to regionprops():

props = regionprops(mask, 'All'); % Or whatever measurements you want.

[Ali Zulfikaroglu]

Thank you really image analyst . I added your function bsxfun and got the result which I want.

Answer 2

[KALYAN ACHARJYA]

Steps:

• Draw
• Right Click (Cursor above the Drawing)
• Create Mask
• Save the Mask

[Ali Zulfikaroglu]

Thank your for helping but I need orginal image extraction in that blue drawed region not binary mask .

[KALYAN ACHARJYA]

After these steps, please follow Image Anayst answer (Below)

[Ali Zulfikaroglu]

Thank for your suggestion also, image analyst answer works