CT segmentation data

[DLinRadiology]

Hi, you might be interested in this 100 CT slice dataset of 60 patients that we have segmented:

http://medicalsegmentation.com/covid19/

Feel free to add to your list if you find it relevant.

[elcronos]

@DLinRadiology Great contribution ! Thanks

[ffkharbat]

Hi, you might be interested in this 100 CT slice dataset of 60 patients that we have segmented:

http://medicalsegmentation.com/covid19/

Feel free to add to your list if you find it relevant.

Thanks a lot. Can we get the original JPG if they are available? I am not familiar with using the nii images, my bad!

[DLinRadiology]

Hi, you might be interested in this 100 CT slice dataset of 60 patients that we have segmented: http://medicalsegmentation.com/covid19/ Feel free to add to your list if you find it relevant.

Thanks a lot. Can we get the original JPG if they are available? I am not familiar with using the nii images, my bad!

The problem is that segmentations were done on the nifti file and it would be a real hassle to convert segmentations back to their original size. The CT images without annotations don't have much value in my opinion and they are not normalized to Hounsfield. Do you still want them?

If you will be working with volumetric medical images (CT or MRI), nifti is a very common file format that you should get familiar with.

[ffkharbat]

Hi, you might be interested in this 100 CT slice dataset of 60 patients that we have segmented: http://medicalsegmentation.com/covid19/ Feel free to add to your list if you find it relevant.

Thanks a lot. Can we get the original JPG if they are available? I am not familiar with using the nii images, my bad!

The problem is that segmentations were done on the nifti file and it would be a real hassle to convert segmentations back to their original size. The CT images without annotations don't have much value in my opinion and they are not normalized to Hounsfield. Do you still want them?

If you will be working with volumetric medical images (CT or MRI), nifti is a very common file format that you should get familiar with.

Thanks again for your response and clarification I am looking into the nifti so that I can use them. Appreciate your help. If you recommend any tutorial for using the filetype into machine learning please let me know. Of course, the site should be cited properly (Y)

[DLinRadiology]

Hi, you might be interested in this 100 CT slice dataset of 60 patients that we have segmented: http://medicalsegmentation.com/covid19/ Feel free to add to your list if you find it relevant.

Thanks a lot. Can we get the original JPG if they are available? I am not familiar with using the nii images, my bad!

The problem is that segmentations were done on the nifti file and it would be a real hassle to convert segmentations back to their original size. The CT images without annotations don't have much value in my opinion and they are not normalized to Hounsfield. Do you still want them? If you will be working with volumetric medical images (CT or MRI), nifti is a very common file format that you should get familiar with.

Thanks again for your response and clarification I am looking into the nifti so that I can use them. Appreciate your help. If you recommend any tutorial for using the filetype into machine learning please let me know. Of course, the site should be cited properly (Y)

People usually use Python for machine/deep learning. In python, the most common library for reading nifti files is nibabel. The usual way to use it is for example:

import nibabel as nib image_volume = nib.load("./tr_im.nii.gz") image_volume_voxel_data = image_volume.get_fdata()

Same for mask data. Then you can do whatever as you have the data loaded into an array.

[sfleisch]

Please look at the NIfTI Image Converter repo. I found it very easy to use.

[jscheithe]

@DLinRadiology thank you very much for uploading your annotation work! Highly appreciated.

Are you also allowed to share the full volumes, including slices that were not annotated? We're building a volume-level classifier, so we can use any confirmed COVID cases.

[DLinRadiology]

@DLinRadiology thank you very much for uploading your annotation work! Highly appreciated.

Are you also allowed to share the full volumes, including slices that were not annotated? We're building a volume-level classifier, so we can use any confirmed COVID cases.

Hi, unfortunately we don't have the full volumes. You can read about how we collected the data here