Skip to content

ivadomed/data-testing

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

15 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

data-testing

Light-weighted data for testing/tutorial

Content

MRI anat

  • sub-unf01: BIDS subject folder with multiple contrast (T2w, T2star and T1w). Original image and derivatives (see below) from ivadomed spine generic example repo were resampled to 1mm isotropic and converted to Float32 to reduce repo size. Subject was randomly selected among the available subjects. Images were oriented according to RPI convention and were cropped with the following bounding box: x=[50, 181], y=[60, 201], z=[10, 26].
  • derivatives/labels/sub-unf01: derivatives folder with the following labels:
    • _seg-manual: spinal cord segmentations for all contrasts,
    • _lesion-manual: dummy labels which are supposed to represent lesion segmentation. WARNING: these are not actual lesion segmentations, but are only here for the purpose of testing the workflow of ivadomed codebase.
    • _labels-disc-manual: dummy label (single voxel), which is supposed to represent disc label. WARNING: this is not an actual label that represent the anatomy, but is only here for the purpose of testing the workflow of ivadomed codebase.
  • bounding_box.json: dictionary to test a specific function from ivadomed/scripts/bounding_box.py.
  • dataset_description.json: this file is needed to described the dataset.
  • df_ref.csv: Used to test a specific function from ivadomed/loader/utils.py.
  • participants.csv: table with subject name and potential metadata.
  • temporary_results.csv: Used to test ivadomed/scripts/compare_models.py.

Microscopy

  • microscopy_png: BIDS dataset folder containing microscopy data in PNG format (BIDS Version 1.7.0)
  • microscopy_png/sub-rat2 and microscopy_png/sub-rat3: BIDS subjects folders with SEM contrast. Original images and derivatives (see below) from AxonDeepSeg data_axondeepseg_sem repo were adapted for testing purposes.
  • microscopy_png/derivatives/labels/: derivatives folder with the following labels:
    • _seg-axon-manual: axon segmentation
    • _seg-myelin-manual: myelin segmentation
  • microscopy_png/df_ref.csv: Used to test a specific function from ivadomed/loader/utils.py.
  • data_test_png_tif: BIDS dataset folder containing microscopy data to test the loading of PNG/TIF, 8/16 bits and Grayscale/RGB/RGBA formats. 256x256 crop extracted and converted to each format from the original image sub-rat8_sample-V915_SEM.png from the AxonDeepSeg data_axondeepseg_sem repo.

CT-scan

  • ct_scan: BIDS subject folder organized according to the file naming and structure of CT-scan BEP024 version 0.0.0 as of 2021-03-22. Original image (Task09_Spleen.tar/imagesTr/spleen2.nii.gz) and derivative (Task09_Spleen.tar/labelsTr/spleen2.nii.gz) from Simpson, A. L., Antonelli, M., Bakas, S., Bilello, M., Farahani, K., Van Ginneken, B., ... & Cardoso, M. J. (2019). A large annotated medical image dataset for the development and evaluation of segmentation algorithms. arXiv preprint arXiv:1902.09063, used under CC-BY-SA 4.0 license. Data available at Medical Segmentation Decathlon. Images were cropped with the following bounding box: x=[(66, 230], y=[114, 280], z=[64, 98].
  • ct_scan/sub-spleen2 : BIDS subject folder with CT-scan contrast (ct).
  • ct_scan/sub-spleen2/derivatives/labels/: derivatives folder with the following label:
    • _seg-manual: spleen segmentation
  • ct_scan/df_ref.csv: Used to test a specific function from ivadomed/loader/utils.py.