How to use graydiffweight() ?

To fuse an image i want to calculate weight for the image. when i use the following line in matlab i am getting error. My image size is c=576,r=1485 My question is: How to calculate weight of the image that will be range from (0-1)? c=67;r=159; w= graydiffweight(s1,c, r,'GrayDifferenceCutoff',25);

Answer 1

[Vidhi Agarwal]

Hi @Rasika Devi,

The error you're encountering in MATLAB might be due to incorrect input parameters or image size mismatches.

The “graydiffweight” function in MATLAB is used to compute weights based on gray-level differences for image segmentation tasks, and it requires specific inputs. Assuming the settings are configured correctly, follow this general method to calculate a weight map for your image:

1. Verify that "s1" is a grayscale image. If it's not, convert it using “rgb2gray” if it's an RGB image.
2. Ensure that the coordinates (c, r) you are using are within the bounds of the image dimensions.
3. GrayDifferenceCutoff: This parameter specifies the maximum gray-level difference for which weights are calculated. Ensure it is appropriate for your image.

Here is sample code that takes into account all of the above:

% Assuming s1 is your input image
if size(s1, 3) == 3
    s1 = rgb2gray(s1); % Convert to grayscale if it's an RGB image
end

% Check the size of the image
[rows, cols] = size(s1);

% Set the coordinates within the image bounds
c = min(max(1, c), cols); % Ensure c is within 1 to cols
r = min(max(1, r), rows); % Ensure r is within 1 to rows

% Calculate the weight map
w = graydiffweight(s1, c, r, 'GrayDifferenceCutoff', 25);

% Normalize the weights to range from 0 to 1
w = w / max(w(:));

% Display the weight map
imshow(w, []);
title('Weight Map');

For detailed understanding of “graydiffweigh” in R2017a, use the following command to access the documentation:

• web(fullfile(docroot, "images/ref/graydiffweight.html"))

Using this command will assist you in accessing the "graydiffweigh" documentation for the currently installed version of MATLAB.

Hope that helps!