Add Curvas dataset

scripts/datasets/medical/check_curvas.py

@@ -0,0 +1,25 @@
+from torch_em.data import MinInstanceSampler
+from torch_em.util.debug import check_loader
+from torch_em.data.datasets.medical import get_curvas_loader
+
+
+ROOT = "/media/anwai/ANWAI/data/curvas"
+
+
+def check_curvas():
+    loader = get_curvas_loader(
+        path=ROOT,
+        patch_shape=(1, 512, 512),
+        batch_size=2,
+        ndim=2,
+        rater="1",
+        resize_inputs=False,
+        download=False,
+        sampler=MinInstanceSampler()
+    )
+
+    check_loader(loader, 8)
+
+
+if __name__ == "__main__":
+    check_curvas()

torch_em/data/datasets/medical/__init__.py

@@ -2,6 +2,7 @@
 from .btcv import get_btcv_dataset, get_btcv_loader
 from .busi import get_busi_dataset, get_busi_loader
 from .camus import get_camus_dataset, get_camus_loader
+from .curvas import get_curvas_dataset, get_curvas_loader
 from .drive import get_drive_dataset, get_drive_loader
 from .papila import get_papila_dataset, get_papila_loader
 from .plethora import get_plethora_dataset, get_plethora_loader

torch_em/data/datasets/medical/curvas.py

@@ -0,0 +1,100 @@
+import os
+from glob import glob
+from natsort import natsorted
+from typing import Tuple, Union
+
+import torch_em
+
+from .. import util
+
+
+URL = "https://zenodo.org/records/11147560/files/training_data.zip"
+CHECKSUM = "02e64b0d963c3a8ac7330c3161f5f76f25ae01a4072bd3032fb1c4048baec2df"
+
+
+def get_curvas_data(path, download):
+    os.makedirs(path, exist_ok=True)
+
+    data_dir = os.path.join(path, "training_set")
+    if os.path.exists(data_dir):
+        return data_dir
+
+    zip_path = os.path.join(path, "training_data.zip")
+    util.download_source(path=zip_path, url=URL, download=download, checksum=CHECKSUM)
+    util.unzip(zip_path=zip_path, dst=path)
+
+    return data_dir
+
+
+def _get_curvas_paths(path, rater, download):
+    data_dir = get_curvas_data(path=path, download=download)
+
+    if not isinstance(rater, list):
+        rater = [str(rater)]
+
+    assert len(rater) == 1, "The segmentations for multiple raters is not supported at the moment."
+
+    image_paths = natsorted(glob(os.path.join(data_dir, "*", "image.nii.gz")))
+    gt_paths = []
+    for _rater in rater:
+        gt_paths.extend(natsorted(glob(os.path.join(data_dir, "*", f"annotation_{_rater}.nii.gz"))))
+
+    return image_paths, gt_paths
+
+
+def get_curvas_dataset(
+    path: Union[os.PathLike, str],
+    patch_shape: Tuple[int, ...],
+    rater: Union[int, list] = ["1"],
+    resize_inputs: bool = False,
+    download: bool = False,
+    **kwargs
+):
+    """Dataset for segmentation of pancreas, kidney and liver in abdominal CT scans.
+
+    NOTE: This dataset has multiple raters annotating the aforementioned organs for all patients.
+
+    The dataset is located at:
+    - https://www.sycaimedical.com/challenge
+    - https://zenodo.org/records/11147560
+
+    Please cite it if you use this dataset for a publication.
+    """
+    image_paths, gt_paths = _get_curvas_paths(path, rater, download)
+
+    if resize_inputs:
+        resize_kwargs = {"patch_shape": patch_shape, "is_rgb": False}
+        kwargs, patch_shape = util.update_kwargs_for_resize_trafo(
+            kwargs=kwargs, patch_shape=patch_shape, resize_inputs=resize_inputs, resize_kwargs=resize_kwargs
+        )
+
+    dataset = torch_em.default_segmentation_dataset(
+        raw_paths=image_paths,
+        raw_key="data",
+        label_paths=gt_paths,
+        label_key="data",
+        patch_shape=patch_shape,
+        **kwargs
+    )
+
+    return dataset
+
+
+def get_curvas_loader(
+    path: Union[os.PathLike, str],
+    patch_shape: Tuple[int, ...],
+    batch_size: int,
+    rater: Union[int, list] = ["1"],
+    resize_inputs: bool = False,
+    download: bool = False,
+    **kwargs
+):
+    """Dataloader for segmentation of pancreas, kidney and liver in abdominal CT scans.
+    See `get_curvas_dataset` for details.
+    """
+    ds_kwargs, loader_kwargs = util.split_kwargs(torch_em.default_segmentation_dataset, **kwargs)
+    dataset = get_curvas_dataset(
+        path=path, patch_shape=patch_shape, rater=rater, resize_inputs=resize_inputs, download=download, **ds_kwargs
+    )
+    loader = torch_em.get_data_loader(dataset=dataset, batch_size=batch_size, **loader_kwargs)
+    return loader