How extract radiomic features from .mat file?

Hi everyone, i'm working on an automatic segmentation algorithm.

I read dicom image and save it in a matlab variable. Once the reading is finished i have a variable with dimension n*n*n_slice.

I segment the target and then i should extract the radiomic features from that file .mat (always with dimension n*n*n_slice).

How can i do it starting from a file matlab (.mat) and not from a real image (i.e. DICOM, nrrd, nifti ....)?

Thanks

[Jan]

By using:

data = load('File.mat')

? This answer sounds too trivial to be useful.

[LORENZO MARZORATI]

No. I don't need to load a mat file. I already have it on my workspace.

I read the dicom image and store it as .mat file. Once i got the target from the previous .mat file, how extract features from THAT .mat file?

I usually use Pyradiomics for features extraction but it works with real image (DICOM ect...). Now i've matlab file, not real image.

I hope to be clear

[Jan]

You do not work with mat-files at all, but with the contents of these files. After load() it is not a file anymore, but an array (or a struct containing arrays). The same happens for dicom files: afert importing them, the data are available as array.

So if you have the array, you cannot distinguish if it came from a mat or dicom file. In both cases you have an array containing the intensity values of pixels. In consequence I do not understand, what the problem is and what you call "real image". If you can apply a method to data of dicom files, you should be able to use exactly the same method to data from other sources. So if you do have a specific problem, explain exactly, which problem it is. It cannot be the format of the file, if the import was successful already.

[LORENZO MARZORATI]

Thanks Jan for the answer.

My aim is extract radiomic features. Usually i have dicom image (CT scan for example) and dicom file of the relative ROI. Having this two (image and ROI), using Pyradiomics, i easly extract radiomic features.

Now i've to do the same things, but before extract features i did a segmentation, creating myself the ROI. In this way, as you said, i've an array (mat file) and no more dicom file.

The question is, How extract radiomic features from the array and no from dicom? Maybe always with Pyradiomics if possible.

Answer 1

[yanqi liu]

sir，may be upload mat file to analysis，use the data as image matrix