The viewer is disabled because this dataset repo requires arbitrary Python code execution. Please consider
removing the
loading script
and relying on
automated data support
(you can use
convert_to_parquet
from the datasets library). If this is not possible, please
open a discussion
for direct help.
PDK - Pachyderm | Determined | KServe
3D Brain MRI Example
Date/Revision: April 30, 2024
This dataset is based on the UCSF-PDGM: The University of California San Francisco Preoperative Diffuse Glioma MRI research dataset, which can be found here:
The original dataset contains data from 495 unique subjects. The dataset is formed by taking several MRI scans for each patient, “skull stripping” the scan (leaving just the brain image), and de-identifying the patient. The result is 4 MRI volumes per subject, as well as a target segmentation mask. In the data folder, you will find a small subset of the data from 87 subjects, which can be used to train a segmentation model.
A sample payload for inference can be found in the sample-payload under 3d-brain.json. Here are the full instructions for how to deploy this on HPE's end-to-end ML platform
- Downloads last month
- 15