[segmentation] Add jsrt database

pyproject.toml

@@ -453,6 +453,9 @@ hrf = "mednet.libs.segmentation.config.data.hrf.default"
 iostar-vessel = "mednet.libs.segmentation.config.data.iostar.vessel"
 iostar-disc = "mednet.libs.segmentation.config.data.iostar.optic_disc"
 
+# jsrt - cxr
+jsrt = "mednet.libs.segmentation.config.data.jsrt.default"
+
 # montgomery county - cxr
 montgomery = "mednet.libs.segmentation.config.data.montgomery.default"
 

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

+# SPDX-FileCopyrightText: Copyright © 2024 Idiap Research Institute <contact@idiap.ch>
+#
+# SPDX-License-Identifier: GPL-3.0-or-later
+"""Japanese Society of Radiological Technology dataset for Lung Segmentation."""
+
+import os
+import pathlib
+
+import numpy as np
+import PIL.Image
+import skimage.exposure
+import torch
+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.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 drishtigs1jsrt dataset."""
+
+    datadir: pathlib.Path
+    """This variable contains the base directory where the database raw data is
+    stored."""
+
+    def __init__(self):
+        self.datadir = pathlib.Path(
+            load_rc().get(CONFIGURATION_KEY_DATADIR, os.path.realpath(os.curdir))
+        )
+
+    def load_pil_raw_12bit_jsrt(self, path: pathlib.Path) -> PIL.Image.Image:
+        """Load a raw 16-bit sample data.
+
+        This method was designed to handle the raw images from the JSRT_ dataset.
+        It reads the data file and applies a simple histogram equalization to the
+        8-bit representation of the image to obtain something along the lines of
+        the PNG (unofficial) version distributed at `JSRT-Kaggle`_.
+
+        Parameters
+        ----------
+        path
+            The full path leading to the image to be loaded.
+
+        Returns
+        -------
+            A PIL image in RGB mode, with `width`x`width` pixels.
+        """
+
+        raw_image = np.fromfile(path, np.dtype(">u2")).reshape(2048, 2048)
+        raw_image[raw_image > 4095] = 4095
+        raw_image = 4095 - raw_image  # invert colors
+        raw_image = (raw_image >> 4).astype(np.uint8)  # 8-bit uint
+        raw_image = skimage.exposure.equalize_hist(raw_image)
+        return PIL.Image.fromarray((raw_image * 255).astype(np.uint8)).convert("RGB")
+
+    def sample(self, sample: tuple[str, str, str]) -> Sample:
+        """Load a single image sample from the disk.
+
+        Parameters
+        ----------
+        sample
+            A tuple containing the path suffix, within the dataset root folder,
+            where to find the image to be loaded, and an integer, representing the
+            sample label.
+
+        Returns
+        -------
+            The sample representation.
+        """
+
+        image = to_tensor(self.load_pil_raw_12bit_jsrt(self.datadir / sample[0]))
+
+        # Combine left and right lung masks into a single tensor
+        target = tv_tensors.Image(
+            to_tensor(
+                np.ma.mask_or(
+                    np.asarray(
+                        PIL.Image.open(self.datadir / sample[1]).convert(
+                            mode="1", dither=None
+                        )
+                    ),
+                    np.asarray(
+                        PIL.Image.open(self.datadir / sample[2]).convert(
+                            mode="1", dither=None
+                        )
+                    ),
+                )
+            ).float()
+        )
+
+        tensor = tv_tensors.Image(image)
+        target = tv_tensors.Image(target)
+        mask = tv_tensors.Mask(torch.ones_like(target))
+
+        return tensor, dict(target=target, mask=mask, name=sample[0])  # type: ignore[arg-type]
+
+
+class DataModule(CachingDataModule):
+    """Japanese Society of Radiological Technology dataset for Lung Segmentation.
+
+    The database includes 154 nodule and 93 non-nodule images.  It contains a total
+    of 247 resolution of 2048 x 2048.  One set of ground-truth lung annotations is
+    available.
+
+    * Reference: [JSRT-2000]_
+    * Original resolution (height x width): 2048 x 2048
+    * Configuration resolution: 1024 x 1024 (after rescaling)
+    * Split reference: [GAAL-2020]_
+    * Protocol ``default``:
+
+    * Training samples: 172 (including labels)
+    * Validation samples: 25 (including labels)
+    * Test samples: 50 (including labels)
+
+    Parameters
+    ----------
+    split_filename
+        Name of the .json file containing the split to load.
+    """
+
+    def __init__(self, split_filename: str):
+        assert __package__ is not None
+        super().__init__(
+            database_split=make_split(__package__, split_filename),
+            raw_data_loader=SegmentationRawDataLoader(),
+            database_name=__package__.rsplit(".", 1)[1],
+            split_name=pathlib.Path(split_filename).stem,
+        )

src/mednet/libs/segmentation/config/data/jsrt/default.py

+# SPDX-FileCopyrightText: Copyright © 2024 Idiap Research Institute <contact@idiap.ch>
+#
+# SPDX-License-Identifier: GPL-3.0-or-later
+"""Japanese Society of Radiological Technology dataset for Lung Segmentation
+(default protocol).
+
+* Split reference: [GAAL-2020]_
+* Configuration resolution: 256 x 256
+* See :py:mod:`deepdraw.data.jsrt` for dataset details
+"""
+
+from mednet.libs.segmentation.config.data.jsrt.datamodule import (
+    DataModule,
+)
+
+datamodule = DataModule("default.json")

src/mednet/libs/segmentation/tests/test_jsrt.py

+# SPDX-FileCopyrightText: Copyright © 2024 Idiap Research Institute <contact@idiap.ch>
+#
+# SPDX-License-Identifier: GPL-3.0-or-later
+"""Tests for jsrt dataset."""
+
+import importlib
+
+import pytest
+from click.testing import CliRunner
+
+
+def id_function(val):
+    if isinstance(val, dict):
+        return str(val)
+    return repr(val)
+
+
+@pytest.mark.parametrize(
+    "split,lengths",
+    [
+        ("default", dict(train=172, validation=25, test=50)),
+    ],
+    ids=id_function,  # just changes how pytest prints it
+)
+def test_protocol_consistency(
+    database_checkers,
+    split: str,
+    lengths: dict[str, int],
+):
+    from mednet.libs.common.data.split import make_split
+
+    database_checkers.check_split(
+        make_split("mednet.libs.segmentation.config.data.jsrt", f"{split}.json"),
+        lengths=lengths,
+    )
+
+
+@pytest.mark.skip_if_rc_var_not_set("datadir.jsrt")
+def test_database_check():
+    from mednet.libs.segmentation.scripts.database import check
+
+    runner = CliRunner()
+    result = runner.invoke(check, ["jsrt"])
+    assert (
+        result.exit_code == 0
+    ), f"Exit code {result.exit_code} != 0 -- Output:\n{result.output}"
+
+
+@pytest.mark.skip_if_rc_var_not_set("datadir.jsrt")
+@pytest.mark.parametrize(
+    "dataset",
+    [
+        "train",
+        "validation",
+        "test",
+    ],
+)
+@pytest.mark.parametrize(
+    "name",
+    [
+        "default",
+    ],
+)
+def test_loading(database_checkers, name: str, dataset: str):
+    datamodule = importlib.import_module(
+        f".{name}",
+        "mednet.libs.segmentation.config.data.jsrt",
+    ).datamodule
+
+    datamodule.model_transforms = []  # should be done before setup()
+    datamodule.setup("predict")  # sets up all datasets
+
+    loader = datamodule.predict_dataloader()[dataset]
+
+    limit = 3  # limit load checking
+    for batch in loader:
+        if limit == 0:
+            break
+        database_checkers.check_loaded_batch(
+            batch,
+            batch_size=1,
+            color_planes=3,
+            expected_num_targets=1,
+        )
+        limit -= 1
+
+
+@pytest.mark.skip_if_rc_var_not_set("datadir.jsrt")
+def test_raw_transforms_image_quality(database_checkers, datadir):
+    reference_histogram_file = str(
+        datadir / "histograms/raw_data/histograms_jsrt_default.json",
+    )
+
+    datamodule = importlib.import_module(
+        ".default",
+        "mednet.libs.segmentation.config.data.jsrt",
+    ).datamodule
+
+    datamodule.model_transforms = []
+    datamodule.setup("predict")
+
+    database_checkers.check_image_quality(datamodule, reference_histogram_file)
+
+
+@pytest.mark.skip_if_rc_var_not_set("datadir.jsrt")
+@pytest.mark.parametrize(
+    "model_name",
+    ["lwnet"],
+)
+def test_model_transforms_image_quality(database_checkers, datadir, model_name):
+    reference_histogram_file = str(
+        datadir / f"histograms/models/histograms_{model_name}_jsrt_default.json",
+    )
+
+    datamodule = importlib.import_module(
+        ".default",
+        "mednet.libs.segmentation.config.data.jsrt",
+    ).datamodule
+
+    model = importlib.import_module(
+        f".{model_name}",
+        "mednet.libs.segmentation.config.models",
+    ).model
+
+    datamodule.model_transforms = model.model_transforms
+    datamodule.setup("predict")
+
+    database_checkers.check_image_quality(
+        datamodule,
+        reference_histogram_file,
+        compare_type="statistical",
+        pearson_coeff_threshold=0.005,
+    )