New "Real World" medical datasets

Please note that this pages lists freshly added new datasets from the medical domain.
This is to emphasize that most currently used datasets in volume rendering do not reflect the current nature of datasets from medical 3D scanners. They are usually too small.

New: The two segmentation files had an inconsistent indexing. If you still use the old ones, you need to add two to every dataset size dimension, e.g., 256x320x128 becomes 258x322x130.
If you use the current ones, you can use the specified size dimensions.




All datasets are binary, storing 8bit/16bit voxels for all slices, for all scanlines, and for all voxels. If 16bit voxels are used, they are stored in Least-Significant-Bit format (LSB), which is commonly used on PC platforms. Note that most medical datasets only use up to 12bits per voxel, so the upper four bits will zero.

Download dataset
(sample image)
Name
Bits per Voxel, Size
Spacing (mm)
Description
Courtesy
mrt_angio.jpg
Head MRT Angiography 16Bits
(10bits set)
416 x 512 x 112
0.412, 0.412, 0.412
3T MRT Time-of-Flight Angiography dataset of a human head. The dataset has been resampled into an isotropic voxel grid (hence the peculiar slice size).
The 8bits downsampled version can be found here.
Özlem Gürvit, Institute for Neuroradiology, Frankfurt, Germany.
mrt_angio.jpg
Head MRT Angiography 16Bits
(10bits set)
256 x 320 x 128
0.66, 0.66, 0.66
3T MRT Time-of-Flight Angiography dataset of a human head. The dataset has been resampled into an isotropic voxel grid (hence the peculiar slice size) and is the same as the one above.
We also provide a segmentation of the arterial blood vessels, where the first (negative) number is the number of all voxels of the segmentation, followed by the voxel array indices of the voxels of the segmentations. Please read the new comment above if you have problems with the voxel indices in the segmenation file.
The 8bits downsampled version can be found here.
Özlem Gürvit, Institute for Neuroradiology, Frankfurt, Germany.
For the segmentation: Dirk Bartz, VCM, University of Tübingen, Germany.
mri_ventricles.jpg
Head MRI CISS 8Bits
(8bits set)
256 x 256 x 124
0.9, 0.9, 0.9
1.5T MRT 3D CISS dataset of a human head that highlights the CSF (Cerebro-Spinal-Fluid) filled cavities of the head.
We also provide a segmentation of the cerebral ventricular system, where the first (negative) number is the number of all voxels of the segmentation, followed by the voxel array indices of the voxels of the segmentations. Please read the new comment above if you have problems with the voxel indices in the segmenation file.
Dirk Bartz, VCM, University of Tübingen, Germany.
stent.jpg
Stented Abdominal Aorta 16Bits
(12bits set)
512 x 512 x 174
0.8398, 0.8398, 3.2
CT Scan of the abdomen and pelvis. The dataset contains also a stent in the abdominal aorta. No contrast agent was used to enhance the blood vessels.
The 8bits downsampled version can be found here.
Michael Meißner, Viatronix Inc., USA.
vertebra.jpg
Head Aneuyrism 16Bits
(12bits set)
512 x 512 x 512
0.1953, 0.1953, 0.1953
Rotational angiography scan of a head with an aneurysm. Only contrasted blood vessels are visible.
The 8bits downsampled version can be found here.
Michael Meißner, Viatronix Inc., USA.
backpack.jpg
Backpack Scan 16Bits
(12bits set)
512 x 512 x 373
0.9766, 0.9766, 1.25
CT scan of a backpack filled with items.
The 8bits downsampled version can be found here.
Kevin Kreeger, Viatronix Inc., USA.
prone.jpg
Colon Prone 16Bits
(12bits set)
512 x 512 x 463
0.625, 0.625, 1.0
CT scan of abdomen in prone orientation (back faces ceiling, belly faces table.
The 8bits downsampled version can be found here.
Walter Reed Army Medical Center, USA. Details can be found here. (Study 289 (213))
supine.jpg
Colon Supine 16Bits
(12bits set)
512 x 512 x 426
0.625, 0.625, 1.0
CT scan of abdomen in supine orientation (back faces table, belly faces ceiling.
The 8bits downsampled version can be found here.
Walter Reed Army Medical Center, USA. Details can be found here. (Study 289 (213))
colon_phantom.jpg
Colon Phantom 16Bits
(12bits set)
512 x 512 x 442
0.9316, 0.9316, 0.5
CT scan of a Colon phantom with several different objects and five pedunculated large polyps in the central object.
The 8bits downsampled version can be found here.
Michael Meißner, Viatronix Inc., USA.


We tried to give credit to the people/organizations that originally provided the respective datasets. In case some references are wrong and should be adjusted, please do not hesitate to send an e-Mail. Thanks!


Last Update: March 21, 2005 Contact: bartz 'at' gris.uni-tuebingen.de