Hugging Face's logo

Datasets:
BiomedSyn
/
CT-eval

Dataset card Files
xet
Community
Dataset Viewer
Duplicate
The dataset viewer is not available for this split.
Cannot load the dataset split (in streaming mode) to extract the first rows.
Error code:   StreamingRowsError
Exception:    ImportError
Message:      To support decoding NIfTI files, please install 'nibabel'.
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/src/worker/utils.py", line 99, in get_rows_or_raise
                  return get_rows(
                         ^^^^^^^^^
                File "/src/libs/libcommon/src/libcommon/utils.py", line 272, in decorator
                  return func(*args, **kwargs)
                         ^^^^^^^^^^^^^^^^^^^^^
                File "/src/services/worker/src/worker/utils.py", line 77, in get_rows
                  rows_plus_one = list(itertools.islice(ds, rows_max_number + 1))
                                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2567, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2103, in __iter__
                  batch = formatter.format_batch(pa_table)
                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/formatting/formatting.py", line 472, in format_batch
                  batch = self.python_features_decoder.decode_batch(batch)
                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/formatting/formatting.py", line 234, in decode_batch
                  return self.features.decode_batch(batch, token_per_repo_id=self.token_per_repo_id) if self.features else batch
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/features/features.py", line 2254, in decode_batch
                  decode_nested_example(self[column_name], value, token_per_repo_id=token_per_repo_id)
                File "/usr/local/lib/python3.12/site-packages/datasets/features/features.py", line 1508, in decode_nested_example
                  return schema.decode_example(obj, token_per_repo_id=token_per_repo_id) if obj is not None else None
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/features/nifti.py", line 172, in decode_example
                  raise ImportError("To support decoding NIfTI files, please install 'nibabel'.")
              ImportError: To support decoding NIfTI files, please install 'nibabel'.

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

YAML Metadata Warning: empty or missing yaml metadata in repo card

Check out the documentation for more information.

CT Generation Evaluation Docker

This Docker container evaluates CT volume generation predictions by running multiple metrics and merging their outputs into a single JSON file.

Metrics

  • FVD_CTNet – Frechet Video Distance computed on 3D CT volumes using the CT-Net backbone.
  • CLIPScore / CLIP_I2I – CLIP-based image-text similarity, reporting both I2T and I2I scores and their mean.
  • FID_2p5D – 2.5D Frechet Inception Distance computed on orthogonal slices (XY, XZ, YZ).

Input Specification

The container accepts either:

# Mount predictions directory or ZIP archive to /input
docker run --rm \
  -v "$(pwd)/input":/input \
  -v "$(pwd)/output":/output \
  forithmus/ctgen-eval:latest

Inside /input, provide either:

  • A set of flattened .mha files under /input or any nested subdirectories.
  • A single .zip archive anywhere under /input containing .mha files.

Notes:

  • The first matching .mha files or the first .zip found will be used for evaluation.

Ground-Truth Data

Ground-truth .mha volumes are baked into the container at:

/opt/app/ground-truth

Each file in this directory should be named by accession or unique identifier and have a .mha extension.

Output Specification

After evaluation, the container writes the merged metrics JSON to /output/metrics.json. The file has the following structure:

{
  "FVD_CTNet": <float>,        // FVD score
  "CLIPScore": <float>,        // CLIP I2T score
  "CLIPScore_I2I": <float>,    // CLIP I2I score
  "CLIPScore_mean": <float>,   // mean CLIP score
  "FID_2p5D_Avg": <float>,     // average 2.5D FID
  "FID_2p5D_XY": <float>,      // XY-plane FID
  "FID_2p5D_XZ": <float>,      // XZ-plane FID
  "FID_2p5D_YZ": <float>       // YZ-plane FID
}

All values are floats rounded to four decimal places.

Testing

To verify functionality, run:

./test.sh

Ensure the script has execute permissions:

chmod +x test.sh

Exporting

Use the export.sh script to set environment variables before running evaluation:

source ./export.sh

This will generate a .tar.gz file for submission to the challenge platform.

For questions or issues, please contact the challenge organizers.
Downloads last month
-