[segmentation] Add montgomery database

ec2ee219 · Daniel CARRON · André Anjos · 7e204e95 · ec2ee219 · ec2ee219
Commit ec2ee219 authored 9 months ago by Daniel CARRON Committed by André Anjos 8 months ago
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -435,7 +435,7 @@ iostar-vessel = "mednet.libs.segmentation.config.data.iostar.vessel"
 iostar-disc = "mednet.libs.segmentation.config.data.iostar.optic_disc"

 # montgomery county - cxr
-# montgomery = "mednet.libs.segmentation.config.data.montgomery.default"
+montgomery = "mednet.libs.segmentation.config.data.montgomery.default"

 # stare dataset - retinography
 stare = "mednet.libs.segmentation.config.data.stare.ah"

--- a/src/mednet/libs/segmentation/config/data/montgomery/__init__.py
+++ b/src/mednet/libs/segmentation/config/data/montgomery/__init__.py
--- a/src/mednet/libs/segmentation/config/data/montgomery/datamodule.py
+++ b/src/mednet/libs/segmentation/config/data/montgomery/datamodule.py
+# SPDX-FileCopyrightText: Copyright © 2024 Idiap Research Institute <contact@idiap.ch>
+#
+# SPDX-License-Identifier: GPL-3.0-or-later
+"""Montgomery DataModule for TB detection."""
+
+import importlib.resources
+import os
+from pathlib import Path
+
+import numpy as np
+import PIL.Image
+import pkg_resources
+import torch
+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 Montgomery 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 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 = PIL.Image.open(Path(self.datadir) / str(sample[0])).convert(
+            mode="RGB"
+        )
+        tensor = tv_tensors.Image(to_tensor(image))
+
+        # 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(
+                            Path(self.datadir) / str(sample[1])
+                        ).convert(mode="1", dither=None)
+                    ),
+                    np.asarray(
+                        PIL.Image.open(
+                            Path(self.datadir) / str(sample[2])
+                        ).convert(mode="1", dither=None)
+                    ),
+                )
+            ).float()
+        )
+
+        mask = tv_tensors.Mask(torch.ones_like(target))
+
+        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 Montgomery 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):
+    """Montgomery DataModule for TB detection.
+
+    The standard digital image database for Tuberculosis was created by the National
+    Library of Medicine, Maryland, USA in collaboration with Shenzhen No.3 People’s
+    Hospital, Guangdong Medical College, Shenzhen, China. The Chest X-rays are from
+
+    * Database reference: [MONTGOMERY-SHENZHEN-2014]_
+    * Original resolution (height x width or width x height): 4020x4892 px or 4892x4020 px
+
+    Data specifications:
+
+    * Raw data input (on disk):
+
+      * PNG images 8 bit grayscale
+      * resolution: fixed to one of the cases above
+
+    * Output image:
+
+      * Transforms:
+
+        * Load raw PNG with :py:mod:`PIL`
+        * Remove black borders
+        * Convert to torch tensor
+        * Torch center cropping to get square image
+
+      * Final specifications
+
+        * Grayscale, encoded as a single plane tensor, 32-bit floats,
+          square at 4020 x 4020 pixels
+        * Labels: 0 (healthy), 1 (active tuberculosis)
+
+    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(),
+        )
--- a/src/mednet/libs/segmentation/config/data/montgomery/default.json
+++ b/src/mednet/libs/segmentation/config/data/montgomery/default.json
+{
+ "train": [
+  [
+   "CXR_png/MCUCXR_0393_1.png",
+   "ManualMask/leftMask/MCUCXR_0393_1.png",
+   "ManualMask/rightMask/MCUCXR_0393_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0294_1.png",
+   "ManualMask/leftMask/MCUCXR_0294_1.png",
+   "ManualMask/rightMask/MCUCXR_0294_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0144_1.png",
+   "ManualMask/leftMask/MCUCXR_0144_1.png",
+   "ManualMask/rightMask/MCUCXR_0144_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0383_1.png",
+   "ManualMask/leftMask/MCUCXR_0383_1.png",
+   "ManualMask/rightMask/MCUCXR_0383_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0309_1.png",
+   "ManualMask/leftMask/MCUCXR_0309_1.png",
+   "ManualMask/rightMask/MCUCXR_0309_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0094_0.png",
+   "ManualMask/leftMask/MCUCXR_0094_0.png",
+   "ManualMask/rightMask/MCUCXR_0094_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0182_1.png",
+   "ManualMask/leftMask/MCUCXR_0182_1.png",
+   "ManualMask/rightMask/MCUCXR_0182_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0091_0.png",
+   "ManualMask/leftMask/MCUCXR_0091_0.png",
+   "ManualMask/rightMask/MCUCXR_0091_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0063_0.png",
+   "ManualMask/leftMask/MCUCXR_0063_0.png",
+   "ManualMask/rightMask/MCUCXR_0063_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0023_0.png",
+   "ManualMask/leftMask/MCUCXR_0023_0.png",
+   "ManualMask/rightMask/MCUCXR_0023_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0083_0.png",
+   "ManualMask/leftMask/MCUCXR_0083_0.png",
+   "ManualMask/rightMask/MCUCXR_0083_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0001_0.png",
+   "ManualMask/leftMask/MCUCXR_0001_0.png",
+   "ManualMask/rightMask/MCUCXR_0001_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0113_1.png",
+   "ManualMask/leftMask/MCUCXR_0113_1.png",
+   "ManualMask/rightMask/MCUCXR_0113_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0048_0.png",
+   "ManualMask/leftMask/MCUCXR_0048_0.png",
+   "ManualMask/rightMask/MCUCXR_0048_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0086_0.png",
+   "ManualMask/leftMask/MCUCXR_0086_0.png",
+   "ManualMask/rightMask/MCUCXR_0086_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0251_1.png",
+   "ManualMask/leftMask/MCUCXR_0251_1.png",
+   "ManualMask/rightMask/MCUCXR_0251_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0027_0.png",
+   "ManualMask/leftMask/MCUCXR_0027_0.png",
+   "ManualMask/rightMask/MCUCXR_0027_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0254_1.png",
+   "ManualMask/leftMask/MCUCXR_0254_1.png",
+   "ManualMask/rightMask/MCUCXR_0254_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0017_0.png",
+   "ManualMask/leftMask/MCUCXR_0017_0.png",
+   "ManualMask/rightMask/MCUCXR_0017_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0015_0.png",
+   "ManualMask/leftMask/MCUCXR_0015_0.png",
+   "ManualMask/rightMask/MCUCXR_0015_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0042_0.png",
+   "ManualMask/leftMask/MCUCXR_0042_0.png",
+   "ManualMask/rightMask/MCUCXR_0042_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0101_0.png",
+   "ManualMask/leftMask/MCUCXR_0101_0.png",
+   "ManualMask/rightMask/MCUCXR_0101_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0043_0.png",
+   "ManualMask/leftMask/MCUCXR_0043_0.png",
+   "ManualMask/rightMask/MCUCXR_0043_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0074_0.png",
+   "ManualMask/leftMask/MCUCXR_0074_0.png",
+   "ManualMask/rightMask/MCUCXR_0074_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0375_1.png",
+   "ManualMask/leftMask/MCUCXR_0375_1.png",
+   "ManualMask/rightMask/MCUCXR_0375_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0099_0.png",
+   "ManualMask/leftMask/MCUCXR_0099_0.png",
+   "ManualMask/rightMask/MCUCXR_0099_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0104_1.png",
+   "ManualMask/leftMask/MCUCXR_0104_1.png",
+   "ManualMask/rightMask/MCUCXR_0104_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0150_1.png",
+   "ManualMask/leftMask/MCUCXR_0150_1.png",
+   "ManualMask/rightMask/MCUCXR_0150_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0102_0.png",
+   "ManualMask/leftMask/MCUCXR_0102_0.png",
+   "ManualMask/rightMask/MCUCXR_0102_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0188_1.png",
+   "ManualMask/leftMask/MCUCXR_0188_1.png",
+   "ManualMask/rightMask/MCUCXR_0188_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0162_1.png",
+   "ManualMask/leftMask/MCUCXR_0162_1.png",
+   "ManualMask/rightMask/MCUCXR_0162_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0195_1.png",
+   "ManualMask/leftMask/MCUCXR_0195_1.png",
+   "ManualMask/rightMask/MCUCXR_0195_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0052_0.png",
+   "ManualMask/leftMask/MCUCXR_0052_0.png",
+   "ManualMask/rightMask/MCUCXR_0052_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0070_0.png",
+   "ManualMask/leftMask/MCUCXR_0070_0.png",
+   "ManualMask/rightMask/MCUCXR_0070_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0372_1.png",
+   "ManualMask/leftMask/MCUCXR_0372_1.png",
+   "ManualMask/rightMask/MCUCXR_0372_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0089_0.png",
+   "ManualMask/leftMask/MCUCXR_0089_0.png",
+   "ManualMask/rightMask/MCUCXR_0089_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0253_1.png",
+   "ManualMask/leftMask/MCUCXR_0253_1.png",
+   "ManualMask/rightMask/MCUCXR_0253_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0092_0.png",
+   "ManualMask/leftMask/MCUCXR_0092_0.png",
+   "ManualMask/rightMask/MCUCXR_0092_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0031_0.png",
+   "ManualMask/leftMask/MCUCXR_0031_0.png",
+   "ManualMask/rightMask/MCUCXR_0031_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0071_0.png",
+   "ManualMask/leftMask/MCUCXR_0071_0.png",
+   "ManualMask/rightMask/MCUCXR_0071_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0228_1.png",
+   "ManualMask/leftMask/MCUCXR_0228_1.png",
+   "ManualMask/rightMask/MCUCXR_0228_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0096_0.png",
+   "ManualMask/leftMask/MCUCXR_0096_0.png",
+   "ManualMask/rightMask/MCUCXR_0096_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0141_1.png",
+   "ManualMask/leftMask/MCUCXR_0141_1.png",
+   "ManualMask/rightMask/MCUCXR_0141_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0057_0.png",
+   "ManualMask/leftMask/MCUCXR_0057_0.png",
+   "ManualMask/rightMask/MCUCXR_0057_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0126_1.png",
+   "ManualMask/leftMask/MCUCXR_0126_1.png",
+   "ManualMask/rightMask/MCUCXR_0126_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0055_0.png",
+   "ManualMask/leftMask/MCUCXR_0055_0.png",
+   "ManualMask/rightMask/MCUCXR_0055_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0005_0.png",
+   "ManualMask/leftMask/MCUCXR_0005_0.png",
+   "ManualMask/rightMask/MCUCXR_0005_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0275_1.png",
+   "ManualMask/leftMask/MCUCXR_0275_1.png",
+   "ManualMask/rightMask/MCUCXR_0275_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0311_1.png",
+   "ManualMask/leftMask/MCUCXR_0311_1.png",
+   "ManualMask/rightMask/MCUCXR_0311_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0026_0.png",
+   "ManualMask/leftMask/MCUCXR_0026_0.png",
+   "ManualMask/rightMask/MCUCXR_0026_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0350_1.png",
+   "ManualMask/leftMask/MCUCXR_0350_1.png",
+   "ManualMask/rightMask/MCUCXR_0350_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0087_0.png",
+   "ManualMask/leftMask/MCUCXR_0087_0.png",
+   "ManualMask/rightMask/MCUCXR_0087_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0035_0.png",
+   "ManualMask/leftMask/MCUCXR_0035_0.png",
+   "ManualMask/rightMask/MCUCXR_0035_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0021_0.png",
+   "ManualMask/leftMask/MCUCXR_0021_0.png",
+   "ManualMask/rightMask/MCUCXR_0021_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0140_1.png",
+   "ManualMask/leftMask/MCUCXR_0140_1.png",
+   "ManualMask/rightMask/MCUCXR_0140_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0058_0.png",
+   "ManualMask/leftMask/MCUCXR_0058_0.png",
+   "ManualMask/rightMask/MCUCXR_0058_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0079_0.png",
+   "ManualMask/leftMask/MCUCXR_0079_0.png",
+   "ManualMask/rightMask/MCUCXR_0079_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0002_0.png",
+   "ManualMask/leftMask/MCUCXR_0002_0.png",
+   "ManualMask/rightMask/MCUCXR_0002_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0061_0.png",
+   "ManualMask/leftMask/MCUCXR_0061_0.png",
+   "ManualMask/rightMask/MCUCXR_0061_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0060_0.png",
+   "ManualMask/leftMask/MCUCXR_0060_0.png",
+   "ManualMask/rightMask/MCUCXR_0060_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0331_1.png",
+   "ManualMask/leftMask/MCUCXR_0331_1.png",
+   "ManualMask/rightMask/MCUCXR_0331_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0108_1.png",
+   "ManualMask/leftMask/MCUCXR_0108_1.png",
+   "ManualMask/rightMask/MCUCXR_0108_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0062_0.png",
+   "ManualMask/leftMask/MCUCXR_0062_0.png",
+   "ManualMask/rightMask/MCUCXR_0062_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0047_0.png",
+   "ManualMask/leftMask/MCUCXR_0047_0.png",
+   "ManualMask/rightMask/MCUCXR_0047_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0080_0.png",
+   "ManualMask/leftMask/MCUCXR_0080_0.png",
+   "ManualMask/rightMask/MCUCXR_0080_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0255_1.png",
+   "ManualMask/leftMask/MCUCXR_0255_1.png",
+   "ManualMask/rightMask/MCUCXR_0255_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0338_1.png",
+   "ManualMask/leftMask/MCUCXR_0338_1.png",
+   "ManualMask/rightMask/MCUCXR_0338_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0038_0.png",
+   "ManualMask/leftMask/MCUCXR_0038_0.png",
+   "ManualMask/rightMask/MCUCXR_0038_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0289_1.png",
+   "ManualMask/leftMask/MCUCXR_0289_1.png",
+   "ManualMask/rightMask/MCUCXR_0289_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0019_0.png",
+   "ManualMask/leftMask/MCUCXR_0019_0.png",
+   "ManualMask/rightMask/MCUCXR_0019_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0029_0.png",
+   "ManualMask/leftMask/MCUCXR_0029_0.png",
+   "ManualMask/rightMask/MCUCXR_0029_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0369_1.png",
+   "ManualMask/leftMask/MCUCXR_0369_1.png",
+   "ManualMask/rightMask/MCUCXR_0369_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0399_1.png",
+   "ManualMask/leftMask/MCUCXR_0399_1.png",
+   "ManualMask/rightMask/MCUCXR_0399_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0173_1.png",
+   "ManualMask/leftMask/MCUCXR_0173_1.png",
+   "ManualMask/rightMask/MCUCXR_0173_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0049_0.png",
+   "ManualMask/leftMask/MCUCXR_0049_0.png",
+   "ManualMask/rightMask/MCUCXR_0049_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0170_1.png",
+   "ManualMask/leftMask/MCUCXR_0170_1.png",
+   "ManualMask/rightMask/MCUCXR_0170_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0051_0.png",
+   "ManualMask/leftMask/MCUCXR_0051_0.png",
+   "ManualMask/rightMask/MCUCXR_0051_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0022_0.png",
+   "ManualMask/leftMask/MCUCXR_0022_0.png",
+   "ManualMask/rightMask/MCUCXR_0022_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0028_0.png",
+   "ManualMask/leftMask/MCUCXR_0028_0.png",
+   "ManualMask/rightMask/MCUCXR_0028_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0100_0.png",
+   "ManualMask/leftMask/MCUCXR_0100_0.png",
+   "ManualMask/rightMask/MCUCXR_0100_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0085_0.png",
+   "ManualMask/leftMask/MCUCXR_0085_0.png",
+   "ManualMask/rightMask/MCUCXR_0085_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0095_0.png",
+   "ManualMask/leftMask/MCUCXR_0095_0.png",
+   "ManualMask/rightMask/MCUCXR_0095_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0348_1.png",
+   "ManualMask/leftMask/MCUCXR_0348_1.png",
+   "ManualMask/rightMask/MCUCXR_0348_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0194_1.png",
+   "ManualMask/leftMask/MCUCXR_0194_1.png",
+   "ManualMask/rightMask/MCUCXR_0194_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0243_1.png",
+   "ManualMask/leftMask/MCUCXR_0243_1.png",
+   "ManualMask/rightMask/MCUCXR_0243_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0097_0.png",
+   "ManualMask/leftMask/MCUCXR_0097_0.png",
+   "ManualMask/rightMask/MCUCXR_0097_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0103_0.png",
+   "ManualMask/leftMask/MCUCXR_0103_0.png",
+   "ManualMask/rightMask/MCUCXR_0103_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0301_1.png",
+   "ManualMask/leftMask/MCUCXR_0301_1.png",
+   "ManualMask/rightMask/MCUCXR_0301_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0213_1.png",
+   "ManualMask/leftMask/MCUCXR_0213_1.png",
+   "ManualMask/rightMask/MCUCXR_0213_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0040_0.png",
+   "ManualMask/leftMask/MCUCXR_0040_0.png",
+   "ManualMask/rightMask/MCUCXR_0040_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0334_1.png",
+   "ManualMask/leftMask/MCUCXR_0334_1.png",
+   "ManualMask/rightMask/MCUCXR_0334_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0142_1.png",
+   "ManualMask/leftMask/MCUCXR_0142_1.png",
+   "ManualMask/rightMask/MCUCXR_0142_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0203_1.png",
+   "ManualMask/leftMask/MCUCXR_0203_1.png",
+   "ManualMask/rightMask/MCUCXR_0203_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0069_0.png",
+   "ManualMask/leftMask/MCUCXR_0069_0.png",
+   "ManualMask/rightMask/MCUCXR_0069_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0013_0.png",
+   "ManualMask/leftMask/MCUCXR_0013_0.png",
+   "ManualMask/rightMask/MCUCXR_0013_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0081_0.png",
+   "ManualMask/leftMask/MCUCXR_0081_0.png",
+   "ManualMask/rightMask/MCUCXR_0081_0.png"
+  ]
+ ],
+ "validation": [
+  [
+   "CXR_png/MCUCXR_0282_1.png",
+   "ManualMask/leftMask/MCUCXR_0282_1.png",
+   "ManualMask/rightMask/MCUCXR_0282_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0068_0.png",
+   "ManualMask/leftMask/MCUCXR_0068_0.png",
+   "ManualMask/rightMask/MCUCXR_0068_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0117_1.png",
+   "ManualMask/leftMask/MCUCXR_0117_1.png",
+   "ManualMask/rightMask/MCUCXR_0117_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0041_0.png",
+   "ManualMask/leftMask/MCUCXR_0041_0.png",
+   "ManualMask/rightMask/MCUCXR_0041_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0352_1.png",
+   "ManualMask/leftMask/MCUCXR_0352_1.png",
+   "ManualMask/rightMask/MCUCXR_0352_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0020_0.png",
+   "ManualMask/leftMask/MCUCXR_0020_0.png",
+   "ManualMask/rightMask/MCUCXR_0020_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0072_0.png",
+   "ManualMask/leftMask/MCUCXR_0072_0.png",
+   "ManualMask/rightMask/MCUCXR_0072_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0059_0.png",
+   "ManualMask/leftMask/MCUCXR_0059_0.png",
+   "ManualMask/rightMask/MCUCXR_0059_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0003_0.png",
+   "ManualMask/leftMask/MCUCXR_0003_0.png",
+   "ManualMask/rightMask/MCUCXR_0003_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0313_1.png",
+   "ManualMask/leftMask/MCUCXR_0313_1.png",
+   "ManualMask/rightMask/MCUCXR_0313_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0223_1.png",
+   "ManualMask/leftMask/MCUCXR_0223_1.png",
+   "ManualMask/rightMask/MCUCXR_0223_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0006_0.png",
+   "ManualMask/leftMask/MCUCXR_0006_0.png",
+   "ManualMask/rightMask/MCUCXR_0006_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0082_0.png",
+   "ManualMask/leftMask/MCUCXR_0082_0.png",
+   "ManualMask/rightMask/MCUCXR_0082_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0046_0.png",
+   "ManualMask/leftMask/MCUCXR_0046_0.png",
+   "ManualMask/rightMask/MCUCXR_0046_0.png"
+  ]
+ ],
+ "test": [
+  [
+   "CXR_png/MCUCXR_0064_0.png",
+   "ManualMask/leftMask/MCUCXR_0064_0.png",
+   "ManualMask/rightMask/MCUCXR_0064_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0218_1.png",
+   "ManualMask/leftMask/MCUCXR_0218_1.png",
+   "ManualMask/rightMask/MCUCXR_0218_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0258_1.png",
+   "ManualMask/leftMask/MCUCXR_0258_1.png",
+   "ManualMask/rightMask/MCUCXR_0258_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0264_1.png",
+   "ManualMask/leftMask/MCUCXR_0264_1.png",
+   "ManualMask/rightMask/MCUCXR_0264_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0362_1.png",
+   "ManualMask/leftMask/MCUCXR_0362_1.png",
+   "ManualMask/rightMask/MCUCXR_0362_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0044_0.png",
+   "ManualMask/leftMask/MCUCXR_0044_0.png",
+   "ManualMask/rightMask/MCUCXR_0044_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0390_1.png",
+   "ManualMask/leftMask/MCUCXR_0390_1.png",
+   "ManualMask/rightMask/MCUCXR_0390_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0354_1.png",
+   "ManualMask/leftMask/MCUCXR_0354_1.png",
+   "ManualMask/rightMask/MCUCXR_0354_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0316_1.png",
+   "ManualMask/leftMask/MCUCXR_0316_1.png",
+   "ManualMask/rightMask/MCUCXR_0316_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0053_0.png",
+   "ManualMask/leftMask/MCUCXR_0053_0.png",
+   "ManualMask/rightMask/MCUCXR_0053_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0166_1.png",
+   "ManualMask/leftMask/MCUCXR_0166_1.png",
+   "ManualMask/rightMask/MCUCXR_0166_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0008_0.png",
+   "ManualMask/leftMask/MCUCXR_0008_0.png",
+   "ManualMask/rightMask/MCUCXR_0008_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0084_0.png",
+   "ManualMask/leftMask/MCUCXR_0084_0.png",
+   "ManualMask/rightMask/MCUCXR_0084_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0004_0.png",
+   "ManualMask/leftMask/MCUCXR_0004_0.png",
+   "ManualMask/rightMask/MCUCXR_0004_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0196_1.png",
+   "ManualMask/leftMask/MCUCXR_0196_1.png",
+   "ManualMask/rightMask/MCUCXR_0196_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0016_0.png",
+   "ManualMask/leftMask/MCUCXR_0016_0.png",
+   "ManualMask/rightMask/MCUCXR_0016_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0075_0.png",
+   "ManualMask/leftMask/MCUCXR_0075_0.png",
+   "ManualMask/rightMask/MCUCXR_0075_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0056_0.png",
+   "ManualMask/leftMask/MCUCXR_0056_0.png",
+   "ManualMask/rightMask/MCUCXR_0056_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0024_0.png",
+   "ManualMask/leftMask/MCUCXR_0024_0.png",
+   "ManualMask/rightMask/MCUCXR_0024_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0367_1.png",
+   "ManualMask/leftMask/MCUCXR_0367_1.png",
+   "ManualMask/rightMask/MCUCXR_0367_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0011_0.png",
+   "ManualMask/leftMask/MCUCXR_0011_0.png",
+   "ManualMask/rightMask/MCUCXR_0011_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0054_0.png",
+   "ManualMask/leftMask/MCUCXR_0054_0.png",
+   "ManualMask/rightMask/MCUCXR_0054_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0090_0.png",
+   "ManualMask/leftMask/MCUCXR_0090_0.png",
+   "ManualMask/rightMask/MCUCXR_0090_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0045_0.png",
+   "ManualMask/leftMask/MCUCXR_0045_0.png",
+   "ManualMask/rightMask/MCUCXR_0045_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0030_0.png",
+   "ManualMask/leftMask/MCUCXR_0030_0.png",
+   "ManualMask/rightMask/MCUCXR_0030_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0077_0.png",
+   "ManualMask/leftMask/MCUCXR_0077_0.png",
+   "ManualMask/rightMask/MCUCXR_0077_0.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0387_1.png",
+   "ManualMask/leftMask/MCUCXR_0387_1.png",
+   "ManualMask/rightMask/MCUCXR_0387_1.png"
+  ],
+  [
+   "CXR_png/MCUCXR_0266_1.png",
+   "ManualMask/leftMask/MCUCXR_0266_1.png",
+   "ManualMask/rightMask/MCUCXR_0266_1.png"
+  ]
+ ]
+}
--- a/src/mednet/libs/segmentation/config/data/montgomery/default.py
+++ b/src/mednet/libs/segmentation/config/data/montgomery/default.py
+# SPDX-FileCopyrightText: Copyright © 2024 Idiap Research Institute <contact@idiap.ch>
+#
+# SPDX-License-Identifier: GPL-3.0-or-later
+"""Montgomery County dataset for Lung Segmentation.
+
+The database includes 58 cases with manifestation of tuberculosis, and 80
+normal cases.  It contains a total of 138 resolution of 4020 x 4892, or
+4892 x 4020. One set of ground-truth lung annotations is available.
+
+* Reference: [MC-2014]_
+* Original resolution (height x width): 4020 x 4892, or 4892 x 4020
+* Configuration resolution: 512 x 512 (after rescaling)
+* Split reference: [GAAL-2020]_
+* Protocol ``default``:
+
+  * Training samples: 96 (including labels)
+  * Validation samples: 14 (including labels)
+  * Test samples: 28 (including labels)
+"""
+
+from mednet.libs.segmentation.config.data.montgomery.datamodule import (
+    DataModule,
+)
+
+datamodule = DataModule("default.json")
--- a/src/mednet/libs/segmentation/tests/data/histograms/models/histograms_lwnet_montgomery_default.json
+++ b/src/mednet/libs/segmentation/tests/data/histograms/models/histograms_lwnet_montgomery_default.json
--- a/src/mednet/libs/segmentation/tests/data/histograms/raw_data/histograms_montgomery_default.json
+++ b/src/mednet/libs/segmentation/tests/data/histograms/raw_data/histograms_montgomery_default.json
--- a/src/mednet/libs/segmentation/tests/test_montgomery.py
+++ b/src/mednet/libs/segmentation/tests/test_montgomery.py
+# SPDX-FileCopyrightText: Copyright © 2024 Idiap Research Institute <contact@idiap.ch>
+#
+# SPDX-License-Identifier: GPL-3.0-or-later
+"""Tests for Montgomery 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=96, validation=14, test=28)),
+    ],
+    ids=id_function,  # just changes how pytest prints it
+)
+def test_protocol_consistency(
+    database_checkers,
+    split: str,
+    lengths: dict[str, int],
+):
+    from mednet.libs.segmentation.config.data.montgomery.datamodule import (
+        make_split,
+    )
+
+    database_checkers.check_split(
+        make_split(f"{split}.json"),
+        lengths=lengths,
+    )
+
+
+@pytest.mark.skip_if_rc_var_not_set("datadir.montgomery")
+def test_database_check():
+    from mednet.libs.segmentation.scripts.database import check
+
+    runner = CliRunner()
+    result = runner.invoke(check, ["montgomery"])
+    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.montgomery")
+@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.montgomery",
+    ).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,
+            prefixes=["CXR_png"],
+        )
+        limit -= 1
+
+
+@pytest.mark.skip_if_rc_var_not_set("datadir.montgomery")
+def test_raw_transforms_image_quality(database_checkers, datadir):
+    reference_histogram_file = str(
+        datadir / "histograms/raw_data/histograms_montgomery_default.json",
+    )
+
+    datamodule = importlib.import_module(
+        ".default",
+        "mednet.libs.segmentation.config.data.montgomery",
+    ).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.montgomery")
+@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}_montgomery_default.json",
+    )
+
+    datamodule = importlib.import_module(
+        ".default",
+        "mednet.libs.segmentation.config.data.montgomery",
+    ).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,
+    )