[segmentation] Add drishtigs1 database

pyproject.toml

@@ -440,6 +440,12 @@ drionsdb-2nd = "mednet.libs.segmentation.config.data.drionsdb.expert2"
 # drive dataset - retinography
 drive = "mednet.libs.segmentation.config.data.drive.default"
 
+# drishti-gs1 - retinography
+drishtigs1-disc-all = "mednet.libs.segmentation.config.data.drishtigs1.optic_disc_all"
+drishtigs1-disc-any = "mednet.libs.segmentation.config.data.drishtigs1.optic_disc_any"
+drishtigs1-cup-all = "mednet.libs.segmentation.config.data.drishtigs1.optic_cup_all"
+drishtigs1-cup-any = "mednet.libs.segmentation.config.data.drishtigs1.optic_cup_any"
+
 # iostar - retinography
 iostar-vessel = "mednet.libs.segmentation.config.data.iostar.vessel"
 iostar-disc = "mednet.libs.segmentation.config.data.iostar.optic_disc"

src/mednet/libs/segmentation/config/data/drishtigs1/datamodule.py

+# SPDX-FileCopyrightText: Copyright © 2024 Idiap Research Institute <contact@idiap.ch>
+#
+# SPDX-License-Identifier: GPL-3.0-or-later
+"""Drishti-GS1 for Optic Disc and Cup Segmentation."""
+
+import os
+import pathlib
+
+import PIL.Image
+import pkg_resources
+from mednet.libs.common.data.datamodule import CachingDataModule
+from mednet.libs.common.data.split import make_split
+from mednet.libs.common.data.typing import Sample
+from mednet.libs.common.models.transforms import crop_image_to_mask
+from mednet.libs.segmentation.data.typing import (
+    SegmentationRawDataLoader as _SegmentationRawDataLoader,
+)
+from torchvision import tv_tensors
+from torchvision.transforms.functional import to_tensor
+
+from ....utils.rc import load_rc
+
+CONFIGURATION_KEY_DATADIR = "datadir." + (__name__.rsplit(".", 2)[-2])
+"""Key to search for in the configuration file for the root directory of this
+database."""
+
+
+class SegmentationRawDataLoader(_SegmentationRawDataLoader):
+    """A specialized raw-data-loader for the drishtigs1 dataset.
+
+    Parameters
+    ----------
+    target_all
+        Indicate whether to use the "all" or "any" target.
+    """
+
+    datadir: pathlib.Path
+    """This variable contains the base directory where the database raw data is
+    stored."""
+
+    def __init__(self, target_all: bool):
+        self.datadir = pathlib.Path(
+            load_rc().get(CONFIGURATION_KEY_DATADIR, os.path.realpath(os.curdir))
+        )
+
+        self._pkg_path = pathlib.Path(
+            pkg_resources.resource_filename(__name__, "masks")
+        )
+
+        self.target_all = target_all
+
+    def sample(self, sample: tuple[str, str, str]) -> Sample:
+        """Load a single image sample from the disk.
+
+        Parameters
+        ----------
+        sample
+            A tuple containing path suffixes to the sample image, target, and mask
+            to be loaded, within the dataset root folder.
+
+        Returns
+        -------
+            The sample representation.
+        """
+
+        image = to_tensor(PIL.Image.open(self.datadir / sample[0]).convert(mode="RGB"))
+
+        if self.target_all:
+            target = to_tensor(
+                PIL.Image.open(self.datadir / sample[1])
+                .convert(mode="RGB", dither=None)
+                .convert("L")
+                .point(lambda p: p > 254, mode="1")
+            )
+        else:
+            target = to_tensor(
+                PIL.Image.open(self.datadir / sample[1])
+                .convert(mode="RGB", dither=None)
+                .convert("L")
+                .point(lambda p: p > 0, mode="1")
+            )
+
+        mask = to_tensor(
+            PIL.Image.open(self._pkg_path / sample[2]).convert(mode="1", dither=None)
+        )
+
+        tensor = tv_tensors.Image(crop_image_to_mask(image, mask))
+        target = tv_tensors.Image(crop_image_to_mask(target, mask))
+        mask = tv_tensors.Mask(crop_image_to_mask(mask, mask))
+
+        return tensor, dict(target=target, mask=mask, name=sample[0])  # type: ignore[arg-type]
+
+
+class DataModule(CachingDataModule):
+    """Drishti-GS1 for Optic Disc and Cup Segmentation.
+
+    Drishti-GS is a dataset meant for validation of segmenting OD, cup and
+    detecting notching.  The images in the Drishti-GS dataset have been collected
+    and annotated by Aravind Eye hospital, Madurai, India. This dataset is of a
+    single population as all subjects whose eye images are part of this dataset are
+    Indians.
+
+    The dataset is divided into two: a training set and a testing set of images.
+    Training images (50) are provided with groundtruths for OD and Cup segmentation
+    and notching information.
+
+    * Reference (including train/test split): [DRISHTIGS1-2014]_
+    * Original resolution (height x width): varying (min: 1749 x 2045, max: 1845 x
+    2468)
+    * Configuration resolution: 1760 x 2048 (after center cropping)
+    * Protocols ``optic-disc`` and ``optic-cup``:
+    * Training: 50
+    * Test: 51
+
+    Parameters
+    ----------
+    split_filename
+        Name of the .json file containing the split to load.
+    target_all
+        Indicate whether to use the "all" or "any" target.
+    """
+
+    def __init__(self, split_filename: str, target_all: bool):
+        assert __package__ is not None
+        super().__init__(
+            database_split=make_split(__package__, split_filename),
+            raw_data_loader=SegmentationRawDataLoader(target_all),
+            database_name=__package__.rsplit(".", 1)[1],
+            split_name=pathlib.Path(split_filename).stem,
+        )