Merge branch 'ground_truth_crop_transform' into 'master'

Ground Truth Box Crop Transform

See merge request biosignal/software/deepdraw!66

pyproject.toml

@@ -171,16 +171,20 @@ combined-disc = "deepdraw.binseg.configs.datasets.combined.od"
 combined-cup = "deepdraw.binseg.configs.datasets.combined.oc"
 # montgomery county - cxr
 montgomery = "deepdraw.binseg.configs.datasets.montgomery.default"
+montgomery-gt = "deepdraw.binseg.configs.datasets.montgomery.default_gtcrop"
 montgomery-xtest = "deepdraw.binseg.configs.datasets.montgomery.xtest"
 # shenzhen - cxr
 shenzhen = "deepdraw.binseg.configs.datasets.shenzhen.default"
 shenzhen-small = "deepdraw.binseg.configs.datasets.shenzhen.default_256"
+shenzhen-gt = "deepdraw.binseg.configs.datasets.shenzhen.default_gtcrop"
 shenzhen-xtest = "deepdraw.binseg.configs.datasets.shenzhen.xtest"
 # jsrt - cxr
 jsrt = "deepdraw.binseg.configs.datasets.jsrt.default"
+jsrt-gt = "deepdraw.binseg.configs.datasets.jsrt.default_gtcrop"
 jsrt-xtest = "deepdraw.binseg.configs.datasets.jsrt.xtest"
 # cxr8 - cxr
 cxr8 = "deepdraw.binseg.configs.datasets.cxr8.default"
+cxr8-gt = "deepdraw.binseg.configs.datasets.cxr8.default_gtcrop"
 cxr8-xtest = "deepdraw.binseg.configs.datasets.cxr8.xtest"
 
 [project.entry-points."detect.config"]

src/deepdraw/binseg/configs/datasets/cxr8/__init__.py

@@ -55,3 +55,51 @@ def _maker_augmented(protocol, n):
             )
         ],
     )
+
+
+def _maker_augmented_gt_box(protocol, n):
+    from .....common.data.transforms import ColorJitter as _jitter
+    from .....common.data.transforms import Compose as _compose
+    from .....common.data.transforms import GaussianBlur as _blur
+    from .....common.data.transforms import GroundTruthCrop as _gtcrop
+    from .....common.data.transforms import RandomHorizontalFlip as _hflip
+    from .....common.data.transforms import RandomRotation as _rotation
+    from .....common.data.transforms import Resize as _resize
+    from ....data.cxr8 import dataset as raw
+    from .. import make_subset
+
+    def mk_aug_subset(subsets, train_transforms, all_transforms):
+        retval = {}
+
+        for key in subsets.keys():
+            retval[key] = make_subset(subsets[key], transforms=all_transforms)
+            if key == "train":
+                retval["__train__"] = make_subset(
+                    subsets[key],
+                    transforms=train_transforms,
+                )
+            else:
+                if key == "validation":
+                    retval["__valid__"] = retval[key]
+
+        if ("__train__" in retval) and ("__valid__" not in retval):
+            retval["__valid__"] = retval["__train__"]
+
+        return retval
+
+    return mk_aug_subset(
+        subsets=raw.subsets(protocol),
+        all_transforms=[_gtcrop(extra_area=0.2), _resize((n, n))],
+        train_transforms=[
+            _compose(
+                [
+                    _gtcrop(extra_area=0.2),
+                    _resize((n, n)),
+                    _rotation(degrees=15, p=0.5),
+                    _hflip(p=0.5),
+                    _jitter(p=0.5),
+                    _blur(p=0.5),
+                ]
+            )
+        ],
+    )

src/deepdraw/binseg/configs/datasets/cxr8/default_gtcrop.py

+# SPDX-FileCopyrightText: Copyright © 2023 Idiap Research Institute <contact@idiap.ch>
+#
+# SPDX-License-Identifier: GPL-3.0-or-later
+
+"""CXR8 Dataset (default protocol)
+
+* Split reference: [GAAL-2020]_
+* Configuration resolution: 256 x 256
+* See :py:mod:`deepdraw.binseg.data.cxr8` for dataset details
+"""
+
+from . import _maker_augmented_gt_box
+
+dataset = _maker_augmented_gt_box("default", 256)

src/deepdraw/binseg/configs/datasets/jsrt/__init__.py

@@ -55,3 +55,51 @@ def _maker_augmented(protocol):
             )
         ],
     )
+
+
+def _maker_augmented_gt_box(protocol):
+    from .....common.data.transforms import ColorJitter as _jitter
+    from .....common.data.transforms import Compose as _compose
+    from .....common.data.transforms import GaussianBlur as _blur
+    from .....common.data.transforms import GroundTruthCrop as _gtcrop
+    from .....common.data.transforms import RandomHorizontalFlip as _hflip
+    from .....common.data.transforms import RandomRotation as _rotation
+    from .....common.data.transforms import Resize as _resize
+    from ....data.jsrt import dataset as raw
+    from .. import make_subset
+
+    def mk_aug_subset(subsets, train_transforms, all_transforms):
+        retval = {}
+
+        for key in subsets.keys():
+            retval[key] = make_subset(subsets[key], transforms=all_transforms)
+            if key == "train":
+                retval["__train__"] = make_subset(
+                    subsets[key],
+                    transforms=train_transforms,
+                )
+            else:
+                if key == "validation":
+                    retval["__valid__"] = retval[key]
+
+        if ("__train__" in retval) and ("__valid__" not in retval):
+            retval["__valid__"] = retval["__train__"]
+
+        return retval
+
+    return mk_aug_subset(
+        subsets=raw.subsets(protocol),
+        all_transforms=[_gtcrop(extra_area=0.2), _resize((256, 256))],
+        train_transforms=[
+            _compose(
+                [
+                    _gtcrop(extra_area=0.2),
+                    _resize((256, 256)),
+                    _rotation(degrees=15, p=0.5),
+                    _hflip(p=0.5),
+                    _jitter(p=0.5),
+                    _blur(p=0.5),
+                ]
+            )
+        ],
+    )

src/deepdraw/binseg/configs/datasets/jsrt/default_gtcrop.py

+# SPDX-FileCopyrightText: Copyright © 2023 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.binseg.data.jsrt` for dataset details
+"""
+
+from . import _maker_augmented_gt_box
+
+dataset = _maker_augmented_gt_box("default")

src/deepdraw/binseg/configs/datasets/montgomery/__init__.py

@@ -57,3 +57,57 @@ def _maker_augmented(protocol):
             )
         ],
     )
+
+
+def _maker_augmented_gt_box(protocol):
+    from .....common.data.transforms import ColorJitter as _jitter
+    from .....common.data.transforms import Compose as _compose
+    from .....common.data.transforms import GaussianBlur as _blur
+    from .....common.data.transforms import GroundTruthCrop as _gtcrop
+    from .....common.data.transforms import RandomHorizontalFlip as _hflip
+    from .....common.data.transforms import RandomRotation as _rotation
+    from .....common.data.transforms import Resize as _resize
+    from .....common.data.transforms import ShrinkIntoSquare as _shrinkintosq
+    from ....data.montgomery import dataset as raw
+    from .. import make_subset
+
+    def mk_aug_subset(subsets, train_transforms, all_transforms):
+        retval = {}
+
+        for key in subsets.keys():
+            retval[key] = make_subset(subsets[key], transforms=all_transforms)
+            if key == "train":
+                retval["__train__"] = make_subset(
+                    subsets[key],
+                    transforms=train_transforms,
+                )
+            else:
+                if key == "validation":
+                    retval["__valid__"] = retval[key]
+
+        if ("__train__" in retval) and ("__valid__" not in retval):
+            retval["__valid__"] = retval["__train__"]
+
+        return retval
+
+    return mk_aug_subset(
+        subsets=raw.subsets(protocol),
+        all_transforms=[
+            _shrinkintosq(),
+            _gtcrop(extra_area=0.2),
+            _resize((256, 256)),
+        ],
+        train_transforms=[
+            _compose(
+                [
+                    _shrinkintosq(),
+                    _gtcrop(extra_area=0.2),
+                    _resize((256, 256)),
+                    _rotation(degrees=15, p=0.5),
+                    _hflip(p=0.5),
+                    _jitter(p=0.5),
+                    _blur(p=0.5),
+                ]
+            )
+        ],
+    )

src/deepdraw/binseg/configs/datasets/montgomery/default_gtcrop.py

+# SPDX-FileCopyrightText: Copyright © 2023 Idiap Research Institute <contact@idiap.ch>
+#
+# SPDX-License-Identifier: GPL-3.0-or-later
+
+"""Montgomery County dataset for Lung Segmentation (default protocol)
+
+* Split reference: [GAAL-2020]_
+* Configuration resolution: 256 x 256
+* See :py:mod:`deepdraw.binseg.data.montgomery` for dataset details
+"""
+
+from . import _maker_augmented_gt_box
+
+dataset = _maker_augmented_gt_box("default")

src/deepdraw/binseg/configs/datasets/shenzhen/__init__.py

@@ -57,3 +57,57 @@ def _maker_augmented(protocol, n):
             )
         ],
     )
+
+
+def _maker_augmented_gt_box(protocol, n):
+    from .....common.data.transforms import ColorJitter as _jitter
+    from .....common.data.transforms import Compose as _compose
+    from .....common.data.transforms import GaussianBlur as _blur
+    from .....common.data.transforms import GroundTruthCrop as _gtcrop
+    from .....common.data.transforms import RandomHorizontalFlip as _hflip
+    from .....common.data.transforms import RandomRotation as _rotation
+    from .....common.data.transforms import Resize as _resize
+    from .....common.data.transforms import ShrinkIntoSquare as _shrinkintosq
+    from ....data.shenzhen import dataset as raw
+    from .. import make_subset
+
+    def mk_aug_subset(subsets, train_transforms, all_transforms):
+        retval = {}
+
+        for key in subsets.keys():
+            retval[key] = make_subset(subsets[key], transforms=all_transforms)
+            if key == "train":
+                retval["__train__"] = make_subset(
+                    subsets[key],
+                    transforms=train_transforms,
+                )
+            else:
+                if key == "validation":
+                    retval["__valid__"] = retval[key]
+
+        if ("__train__" in retval) and ("__valid__" not in retval):
+            retval["__valid__"] = retval["__train__"]
+
+        return retval
+
+    return mk_aug_subset(
+        subsets=raw.subsets(protocol),
+        all_transforms=[
+            _shrinkintosq(),
+            _gtcrop(extra_area=0.2),
+            _resize((n, n)),
+        ],
+        train_transforms=[
+            _compose(
+                [
+                    _shrinkintosq(),
+                    _gtcrop(extra_area=0.2),
+                    _resize((n, n)),
+                    _rotation(degrees=15, p=0.5),
+                    _hflip(p=0.5),
+                    _jitter(p=0.5),
+                    _blur(p=0.5),
+                ]
+            )
+        ],
+    )

src/deepdraw/binseg/configs/datasets/shenzhen/default_gtcrop.py

+# SPDX-FileCopyrightText: Copyright © 2023 Idiap Research Institute <contact@idiap.ch>
+#
+# SPDX-License-Identifier: GPL-3.0-or-later
+
+"""Shenzhen dataset for Lung Segmentation (default protocol)
+
+* Split reference: [GAAL-2020]_
+* Configuration resolution: 256 x 256
+* See :py:mod:`deepdraw.binseg.data.shenzhen` for dataset details
+"""
+
+from . import _maker_augmented_gt_box
+
+dataset = _maker_augmented_gt_box("default", 256)

src/deepdraw/common/data/transforms.py

@@ -430,3 +430,110 @@ class GetBoundingBox:
         target["labels"] = labels
 
         return [args[self.image], target]
+
+
+class GroundTruthCrop:
+    """Crop image in a square keeping only the area with the ground truth.
+
+    This transform can crop all images given a ground-truth mask as reference.
+    Notice that the crop will result in a square image at the end, which means
+    that it will keep the bigger dimension and adjust the smaller one to fit
+    into a square. There's an option to add extra area around the gt bounding
+    box. If resulting dimensions are larger than the boundaries of the image,
+    minimal padding will be done to keep the image in a square shape.
+
+    Parameters
+    ----------
+
+    reference : :py:class:`int`, Optional
+        Which reference part of the sample to use for getting coordinates.
+        If not set, use the second object on the sample (typically, the mask).
+
+    extra_area : :py:class:`float`, Optional
+        Multiplier that will add the extra area around the ground-truth
+        bounding box. Example: 0.1 will result in a crop with dimensions of
+        the largest side increased by 10%. If not set, the default will be 0
+        (only the ground-truth box).
+    """
+
+    def __init__(self, reference=1, extra_area=0.0):
+        self.reference = reference
+        self.extra_area = extra_area
+
+    def __call__(self, *args):
+        ref = args[self.reference]
+
+        max_w, max_h = ref.size
+
+        where = numpy.where(ref)
+        y0 = numpy.min(where[0])
+        y1 = numpy.max(where[0])
+        x0 = numpy.min(where[1])
+        x1 = numpy.max(where[1])
+
+        w = x1 - x0
+        h = y1 - y0
+
+        extra_x = self.extra_area * w / 2
+        extra_y = self.extra_area * h / 2
+
+        new_w = (1 + self.extra_area) * w
+        new_h = (1 + self.extra_area) * h
+
+        diff = abs(new_w - new_h) / 2
+
+        if new_w == new_h:
+            x0_new = x0.copy() - extra_x
+            x1_new = x1.copy() + extra_x
+            y0_new = y0.copy() - extra_y
+            y1_new = y1.copy() + extra_y
+
+        elif new_w > new_h:
+            x0_new = x0.copy() - extra_x
+            x1_new = x1.copy() + extra_x
+            y0_new = y0.copy() - extra_y - diff
+            y1_new = y1.copy() + extra_y + diff
+
+        else:
+            x0_new = x0.copy() - extra_x - diff
+            x1_new = x1.copy() + extra_x + diff
+            y0_new = y0.copy() - extra_y
+            y1_new = y1.copy() + extra_y
+
+        border = (x0_new, y0_new, max_w - x1_new, max_h - y1_new)
+
+        def _expand_img(
+            pil_img, background_color, x0_pad=0, x1_pad=0, y0_pad=0, y1_pad=0
+        ):
+            width = pil_img.size[0] + x0_pad + x1_pad
+            height = pil_img.size[1] + y0_pad + y1_pad
+
+            result = PIL.Image.new(
+                pil_img.mode, (width, height), background_color
+            )
+            result.paste(pil_img, (x0_pad, y0_pad))
+            return result
+
+        def _black_background(i):
+            return (0, 0, 0) if i.mode == "RGB" else 0
+
+        d_x0 = numpy.rint(max([0 - x0_new, 0])).astype(int)
+        d_y0 = numpy.rint(max([0 - y0_new, 0])).astype(int)
+        d_x1 = numpy.rint(max([x1_new - max_w, 0])).astype(int)
+        d_y1 = numpy.rint(max([y1_new - max_h, 0])).astype(int)
+
+        new_args = [
+            _expand_img(
+                k,
+                _black_background(k),
+                x0_pad=d_x0,
+                x1_pad=d_x1,
+                y0_pad=d_y0,
+                y1_pad=d_y1,
+            )
+            for k in args
+        ]
+
+        new_args = [PIL.ImageOps.crop(k, border) for k in new_args]
+
+        return new_args