[segmentation] Add chasedb1 database

pyproject.toml

@@ -423,8 +423,14 @@ visceral = "mednet.config.data.visceral.default"
 
 lwnet = "mednet.libs.segmentation.config.models.lwnet"
 
-# drive dataset - retinal vessel segmentation
+# chase-db1 - retinography
+chasedb1 = "mednet.libs.segmentation.config.data.chasedb1.first_annotator"
+chasedb1-2nd = "mednet.libs.segmentation.config.data.chasedb1.second_annotator"
+
+# drive dataset - retinography
 drive = "mednet.libs.segmentation.config.data.drive.default"
+
+# stare dataset - retinography
 stare = "mednet.libs.segmentation.config.data.stare.ah"
 stare-2nd = "mednet.libs.segmentation.config.data.stare.vk"
 

src/mednet/libs/segmentation/config/data/chasedb1/datamodule.py

+# SPDX-FileCopyrightText: Copyright © 2024 Idiap Research Institute <contact@idiap.ch>
+#
+# SPDX-License-Identifier: GPL-3.0-or-later
+"""CHASE-DB1 dataset for Vessel Segmentation."""
+
+import importlib.resources
+import os
+from pathlib import Path
+
+import PIL.Image
+import pkg_resources
+from mednet.libs.common.data.datamodule import CachingDataModule
+from mednet.libs.common.data.split import JSONDatabaseSplit
+from mednet.libs.common.data.typing import DatabaseSplit, Sample
+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 Chase-db1 dataset."""
+
+    datadir: str
+    """This variable contains the base directory where the database raw data is
+    stored."""
+
+    def __init__(self):
+        self.datadir = load_rc().get(
+            CONFIGURATION_KEY_DATADIR, os.path.realpath(os.curdir)
+        )
+
+        self._pkg_path = pkg_resources.resource_filename(__name__, "masks")
+
+    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 = PIL.Image.open(Path(self.datadir) / str(sample[0])).convert(
+            mode="RGB"
+        )
+        tensor = tv_tensors.Image(to_tensor(image))
+        target = tv_tensors.Image(
+            to_tensor(
+                PIL.Image.open(Path(self.datadir) / str(sample[1])).convert(
+                    mode="1", dither=None
+                )
+            )
+        )
+        mask = tv_tensors.Mask(
+            to_tensor(
+                PIL.Image.open(Path(self._pkg_path) / str(sample[2])).convert(
+                    mode="1", dither=None
+                )
+            )
+        )
+
+        return tensor, dict(target=target, mask=mask, name=sample[0])  # type: ignore[arg-type]
+
+
+def make_split(basename: str) -> DatabaseSplit:
+    """Return a database split for the Chase-db1 database.
+
+    Parameters
+    ----------
+    basename
+        Name of the .json file containing the split to load.
+
+    Returns
+    -------
+        An instance of DatabaseSplit.
+    """
+
+    return JSONDatabaseSplit(
+        importlib.resources.files(__name__.rsplit(".", 1)[0]).joinpath(basename)
+    )
+
+
+class DataModule(CachingDataModule):
+    """CHASE-DB1 dataset for Vessel Segmentation.
+
+    The CHASE_DB1 is a retinal vessel reference dataset acquired from multiethnic
+    school children. This database is a part of the Child Heart and Health Study in
+    England (CHASE), a cardiovascular health survey in 200 primary schools in
+    London, Birmingham, and Leicester. The ocular imaging was carried out in
+    46 schools and demonstrated associations between retinal vessel tortuosity and
+    early risk factors for cardiovascular disease in over 1000 British primary
+    school children of different ethnic origin. The retinal images of both of the
+    eyes of each child were recorded with a hand-held Nidek NM-200-D fundus camera.
+    The images were captured at 30 degrees FOV camera. The dataset of images are
+    characterized by having nonuniform back-ground illumination, poor contrast of
+    blood vessels as compared with the background and wider arteriolars that have a
+    bright strip running down the centre known as the central vessel reflex.
+
+    * Reference: [CHASEDB1-2012]_
+    * Original resolution (height x width): 960 x 999
+    * Split reference: [CHASEDB1-2012]_
+    * Protocol ``first-annotator``:
+
+    * Training samples: 8 (including labels from annotator "1stHO")
+    * Test samples: 20 (including labels from annotator "1stHO")
+
+    * Protocol ``second-annotator``:
+
+    * Training samples: 8 (including labels from annotator "2ndHO")
+    * Test samples: 20 (including labels from annotator "2ndHO")
+
+    Parameters
+    ----------
+    split_filename
+        Name of the .json file containing the split to load.
+    """
+
+    def __init__(self, split_filename: str):
+        super().__init__(
+            database_split=make_split(split_filename),
+            raw_data_loader=SegmentationRawDataLoader(),
+        )

src/mednet/libs/segmentation/config/data/chasedb1/first-annotator.json

+{
+ "train": [
+  [
+   "Image_11L.jpg",
+   "Image_11L_1stHO.png",
+   "Image_11L.png"
+  ],
+  [
+   "Image_11R.jpg",
+   "Image_11R_1stHO.png",
+   "Image_11R.png"
+  ],
+  [
+   "Image_12L.jpg",
+   "Image_12L_1stHO.png",
+   "Image_12L.png"
+  ],
+  [
+   "Image_12R.jpg",
+   "Image_12R_1stHO.png",
+   "Image_12R.png"
+  ],
+  [
+   "Image_13L.jpg",
+   "Image_13L_1stHO.png",
+   "Image_13L.png"
+  ],
+  [
+   "Image_13R.jpg",
+   "Image_13R_1stHO.png",
+   "Image_13R.png"
+  ],
+  [
+   "Image_14L.jpg",
+   "Image_14L_1stHO.png",
+   "Image_14L.png"
+  ],
+  [
+   "Image_14R.jpg",
+   "Image_14R_1stHO.png",
+   "Image_14R.png"
+  ]
+ ],
+ "test": [
+  [
+   "Image_01L.jpg",
+   "Image_01L_1stHO.png",
+   "Image_01L.png"
+  ],
+  [
+   "Image_01R.jpg",
+   "Image_01R_1stHO.png",
+   "Image_01R.png"
+  ],
+  [
+   "Image_02L.jpg",
+   "Image_02L_1stHO.png",
+   "Image_02L.png"
+  ],
+  [
+   "Image_02R.jpg",
+   "Image_02R_1stHO.png",
+   "Image_02R.png"
+  ],
+  [
+   "Image_03L.jpg",
+   "Image_03L_1stHO.png",
+   "Image_03L.png"
+  ],
+  [
+   "Image_03R.jpg",
+   "Image_03R_1stHO.png",
+   "Image_03R.png"
+  ],
+  [
+   "Image_04L.jpg",
+   "Image_04L_1stHO.png",
+   "Image_04L.png"
+  ],
+  [
+   "Image_04R.jpg",
+   "Image_04R_1stHO.png",
+   "Image_04R.png"
+  ],
+  [
+   "Image_05L.jpg",
+   "Image_05L_1stHO.png",
+   "Image_05L.png"
+  ],
+  [
+   "Image_05R.jpg",
+   "Image_05R_1stHO.png",
+   "Image_05R.png"
+  ],
+  [
+   "Image_06L.jpg",
+   "Image_06L_1stHO.png",
+   "Image_06L.png"
+  ],
+  [
+   "Image_06R.jpg",
+   "Image_06R_1stHO.png",
+   "Image_06R.png"
+  ],
+  [
+   "Image_07L.jpg",
+   "Image_07L_1stHO.png",
+   "Image_07L.png"
+  ],
+  [
+   "Image_07R.jpg",
+   "Image_07R_1stHO.png",
+   "Image_07R.png"
+  ],
+  [
+   "Image_08L.jpg",
+   "Image_08L_1stHO.png",
+   "Image_08L.png"
+  ],
+  [
+   "Image_08R.jpg",
+   "Image_08R_1stHO.png",
+   "Image_08R.png"
+  ],
+  [
+   "Image_09L.jpg",
+   "Image_09L_1stHO.png",
+   "Image_09L.png"
+  ],
+  [
+   "Image_09R.jpg",
+   "Image_09R_1stHO.png",
+   "Image_09R.png"
+  ],
+  [
+   "Image_10L.jpg",
+   "Image_10L_1stHO.png",
+   "Image_10L.png"
+  ],
+  [
+   "Image_10R.jpg",
+   "Image_10R_1stHO.png",
+   "Image_10R.png"
+  ]
+ ]
+}

src/mednet/libs/segmentation/config/data/chasedb1/first_annotator.py

+# SPDX-FileCopyrightText: Copyright © 2024 Idiap Research Institute <contact@idiap.ch>
+#
+# SPDX-License-Identifier: GPL-3.0-or-later
+"""CHASE-DB1 dataset for Vessel Segmentation."""
+
+from mednet.libs.segmentation.config.data.chasedb1.datamodule import DataModule
+
+datamodule = DataModule("first-annotator.json")