How to calculate variance, kurtosis, and skewness of a ROI (Region of Interest) in grayscale image?

I'm going to calculate means, standard deviation, kurtosis, skewness, and also variance from a ROI of grayscale image. I create select the ROI using mask. This is the code I use

Img = imread('image_name.jpg');
GrayImg = rgb2gray(Img);
h = impoly(gca);
wait(h);
mask = createMask(h);
Mean = mean2(GrayImg(mask));
StandardDeviation = std2(GrayImg(mask));

How to calculate variance, kurtosis and skewness of the grayscale image using the mask? Is the code to calculate mean and standard deviation is correct?

Answer 1

[Image Analyst]

See my attached demo for computing image moments. Adapt as needed.

Answer 2

[ag]

Hi Yogi,

To calculate the statistical measures such as mean, standard deviation, variance, kurtosis, and skewness for a region of interest (ROI) in a grayscale image, you can utilize a mask to define the ROI. Below is an improved version of your code, which accurately performs these calculations:

% Read the image and convert it to grayscale
Img = imread('image_name.jpg');
GrayImg = rgb2gray(Img);

% Select the ROI using a polygon tool
h = impoly(gca);
wait(h);

% Generate a mask from the selected polygon
mask = createMask(h);

% Extract the pixel values within the mask
maskedPixels = GrayImg(mask);

% Compute the statistical measures
Mean = mean(maskedPixels);
StandardDeviation = std(maskedPixels);
Variance = var(maskedPixels);
Kurtosis = kurtosis(double(maskedPixels)); % Convert to double for kurtosis
Skewness = skewness(double(maskedPixels)); % Convert to double for skewness

% Display the results

For more details, please refer to the following MathWorks documentations:

• mean - https://www.mathworks.com/help/matlab/ref/double.mean.html
• std - https://www.mathworks.com/help/matlab/ref/double.std.html
• var - https://www.mathworks.com/help/matlab/ref/double.var.html
• kurtosis - https://www.mathworks.com/help/stats/kurtosis.html
• skewness - https://www.mathworks.com/help/stats/skewness.html

Hope this helps!