convert .npy segmentation file into .nrrd

[silkezimm]

Dear all,

i have the problem that i have DICOM images and a npy file of segmented data as input to pyradiomics. #
i converted the DICOM images in nrrd format with 3Dslicer. now i need to convert the segmentation/label file .npy into .nrrd.

i cant solve this;> can anyone help me and provide a solution?

thank you so much,
Silke

[ivanzhovannik]

Hi @silkezimm,

In order to give you good advice, could you please share more info about our .npy file? Given file name filename.npy (change to your file name), please run and post below:

import numpy as np

mask_array  = np.load("filename.npy", allow_pickle=True)
print(f"Shape of my mask array is {mask_array.shape}")
print(f"Unique mask values are {np.unique(mask_array)}")

Usually .npy files contain only the voxel intensities and not the grid, which can be used to understand how to place these voxel intensities in the space.

On the other hand, image files such as .nrrd contain both the grid and voxel intensities. Given your image file image.nrrd (change) please run and post below:

import SimpleITK as sitk

image = sitk.ReadImage("image.nrrd")
print(f"Image size {sitk.GetSize(image)}, image spacing and origin: {sitk.GetSpacing(image)}, {sitk.GetOrigin(image)}")

[silkezimm]

HI! i solved this issue, thank you soo much for your kind reply :) actually i have another issue with using pyradiomics.. i was wondering wheter you can help me.. i am using pyradiomics from the command line. i would like to know whether i extracted the features without param file or any specification from the original image or the segmented region (with the mask file..). i am performing batch feature extraction with the command : pyradiomics <path/to/input> , this works nicely, now i have created a parameter file in .yaml format, and it says on the readthedocs homepage to use the --param argument in the command line. here i struggle as to how exactly tell the software to use my specific file. can you help me here and provide the exact command i shall write? i hope you can help.. best silke Am So., 30. Juli 2023 um 14:03 Uhr schrieb Ivan Ye. Zhovannik < ***@***.***>:

…

Hi @silkezimm <https://github.com/silkezimm>, In order to give you good advice, could you please share more info about our .npy file? Given file name filename.npy (*change to your file name*), please run and post below: import numpy as np mask_array = np.load("filename.npy", allow_pickle=True) print(f"Shape of my mask array is {mask_array.shape}") print(f"Unique mask values are {np.unique(mask_array)}") Usually .npy files contain only the voxel intensities and not the grid, which can be used to understand how to place these voxel intensities in the space. On the other hand, image files such as .nrrd contain both the grid and voxel intensities. Given your image file image.nrrd (*change*) please run and post below: import SimpleITK as sitk image = sitk.ReadImage("image.nrrd") print(f"Image size {sitk.GetSize(image)}, image spacing and origin: {sitk.GetSpacing(image)}, {sitk.GetOrigin(image)}") — Reply to this email directly, view it on GitHub <#12 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AXAMGXKTJWBSSVCJM6N6LWLXSZESBANCNFSM6AAAAAA25EDBVM> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[ivanzhovannik]

@silkezimm like any other command in cmd/bash/other shell, I guess. See the usage here: https://pyradiomics.readthedocs.io/en/latest/usage.html
See more in Python’s argparse package how to pass flags such as —params

[silkezimm]

Yes....I have read the documentation and still can't understand the --param .... Sorry....

…

On Thu, 3 Aug 2023, 19:58 Ivan Ye. Zhovannik, ***@***.***> wrote: @silkezimm <https://github.com/silkezimm> like any other command in cmd/bash/other shell, I guess. See the usage here: https://pyradiomics.readthedocs.io/en/latest/usage.html See more in Python’s argparse package how to pass flags such as —params — Reply to this email directly, view it on GitHub <#12 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AXAMGXLUWDSLFHQRBYMH24DXTPRDLANCNFSM6AAAAAA25EDBVM> . You are receiving this because you were mentioned.Message ID: ***@***.***>