I work on MATLAB to characterise texture for a region of interest (ROI) in my images. I am masking the ROI from the image, and I want to apply my image analysis methods on the ROI only. The problem is that MATLAB is vectorising the ROI every time I do the following to calculate the fractional dimension (FD) for example:
frac_dim = FD(image(mask))
As a result, do you have any recommendations so that MATLAB can take into consideration the ROI while conserving the image aspect of it in order to preserve the spatial aspect?
Thank you for your help,
Vera
If you only want the image data within the ROI then create a new variable ROIImg = image(roiRowStart:roiRowEnd,roiColStart:roiColEnd);
Is it your method of defining 'mask' that is vactorising the data? Or is it the FD function? I can't find reference of FD being a built-in Matlab function so I can't look at the code I'm afraid.
In all the cases, you will have the constraint that your image has to be rectangular. To get a ROI, you can get is through imcrop(). To find the minimum rectangle around the polygon define inside the mask image, you can use this code: http://www.mathworks.com/matlabcentral/fileexchange/34767-a-suite-of-minimal-bounding-objects and then affect to zero the pixel which are not of interest.
It should be enough for computing the fractal dimension.
Hi Vera,
in cases where your mask is not rectangular you can use the function find, to get the indexes of the pixels inside your mask.
ids = find(mask>0); % for example
Once you have this you only have to perform the analysis over the pixels/regions/or the whole set of masked pixels.
If you want to use patches over a pixel, by using the function
sub2ind(imageSize, ids(2), ids(1));
I hope it helps.
Hi Vera,
I posted a Matlab function on Matlab file exchange that extracts image data from an ROI that you can draw. Alternatively you can define a path in x and y coordinates (in whatever unit you specified when you draw the image with imagesc, which per default is pixel numbers). The function will close the path and extract the image data and give some statistics as well (mean stdev median ROI area)
Hope this works for you.
plz check the below link
http://www.mathworks.com/matlabcentral/fileexchange/34767-a-suite-of-minimal
Thank you all for your very informative answers. You have been a great help!
https://in.mathworks.com/matlabcentral/fileexchange/34767-a-suite-of-minimal-bounding-objects
please see this link
- Badges
- Science topic
- Similar topics
- Geoscience
- Remote Sensing