From 72cc30398bbc3cc5b3de7a804d15910d67bd9e51 Mon Sep 17 00:00:00 2001
From: Gregory Lee <grlee77@gmail.com>
Date: Tue, 11 Jul 2023 15:36:39 -0400
Subject: [PATCH] Sync cuCIM API with scikit-image 0.21 (#573)

# Description

This MR ports several recent changes implemented upstream in scikit-image 0.21rc0. Some are just minor refactoring or documentation fixes, but I have commented on the more significant changes below so we can reuse that for the release note.

If preferred, I can split the items below into separate small PRs.

## API Changes / Deprecations
- The `random_state` argument of `medial_axis` and `unsupervised_wiener` is now deprecated and will be removed in the future. The new argument name, `seed`, should be used instead.

- The formerly private function `_invariant_denoise` has been renamed `denoise_invariant` and is now part of the public `cucim.skimage.restoration` API.

- The existing function `cucim.skimage.color.get_xyz_coords` has been renamed `cucim.skimage.color.xyz_tristimulus_values`. The former function name is deprecated and will be removed in the future.

## Enhancements
- A new `return_mapping` option was added to `join_segmentations`. This provides an additional output with a mapping between the labels in the joined segmentation and the original ones.

- Added support for y-axis shear to the [2D AffineTransform](https://github.com/scikit-image/scikit-image/pull/6752)

## Backwards Incompatible Changes
- A bug fix to the energy calculation in `chan_vese` may result in different output from previous versions.

## Bug Fixes
- Fixed [bug in energy calculation in Chan-Vese](https://github.com/scikit-image/scikit-image/pull/6902)

- Fixed a [bug in 2D shear calculations for `AffineTransform`](https://github.com/scikit-image/scikit-image/pull/6717)

Authors:
  - Gregory Lee (https://github.com/grlee77)
  - https://github.com/jakirkham

Approvers:
  - Gigon Bae (https://github.com/gigony)

URL: https://github.com/rapidsai/cucim/pull/573
---
 .../cucim/src/cucim/skimage/_shared/utils.py  | 139 ++++++++----
 .../cucim/src/cucim/skimage/color/__init__.py |   6 +-
 .../src/cucim/skimage/color/colorconv.py      | 203 ++++++++++++++----
 .../cucim/src/cucim/skimage/color/delta_e.py  |   4 +-
 .../skimage/color/tests/test_colorconv.py     |  17 +-
 .../src/cucim/skimage/exposure/_adapthist.py  |  11 +-
 .../src/cucim/skimage/exposure/exposure.py    |  14 +-
 .../skimage/exposure/tests/test_exposure.py   |   9 +-
 .../src/cucim/skimage/feature/__init__.py     |   1 -
 .../cucim/src/cucim/skimage/feature/_canny.py |   3 +
 .../cucim/skimage/feature/tests/test_canny.py |  23 +-
 .../src/cucim/skimage/filters/__init__.py     |  18 +-
 .../src/cucim/skimage/filters/lpi_filter.py   |  65 +++---
 .../cucim/src/cucim/skimage/filters/ridges.py |   4 +-
 .../src/cucim/skimage/filters/thresholding.py |   7 +
 .../cucim/skimage/morphology/_skeletonize.py  |  24 ++-
 .../src/cucim/skimage/restoration/__init__.py |   3 +-
 .../skimage/restoration/deconvolution.py      |  21 +-
 .../cucim/skimage/restoration/j_invariant.py  |  42 +++-
 .../restoration/tests/test_j_invariant.py     |  20 +-
 .../restoration/tests/test_restoration.py     |  19 +-
 .../cucim/skimage/segmentation/_chan_vese.py  |  83 +++----
 .../src/cucim/skimage/segmentation/_join.py   |  29 ++-
 .../skimage/segmentation/tests/test_join.py   |   6 +
 .../src/cucim/skimage/transform/_geometric.py | 175 ++++++++++++---
 .../src/cucim/skimage/transform/_warps.py     |  18 +-
 .../skimage/transform/tests/test_geometric.py |  32 +++
 .../skimage/transform/tests/test_warps.py     |   4 +-
 28 files changed, 726 insertions(+), 274 deletions(-)

diff --git a/python/cucim/src/cucim/skimage/_shared/utils.py b/python/cucim/src/cucim/skimage/_shared/utils.py
index b4371ff1b..51b25e198 100644
--- a/python/cucim/src/cucim/skimage/_shared/utils.py
+++ b/python/cucim/src/cucim/skimage/_shared/utils.py
@@ -9,17 +9,48 @@
 
 from ._warnings import all_warnings, warn  # noqa
 
+__all__ = ['deprecate_func', 'get_bound_method_class', 'all_warnings',
+           'safe_as_int', 'check_shape_equality', 'check_nD', 'warn',
+           'reshape_nd', 'identity', 'slice_at_axis']
 
-class skimage_deprecation(Warning):
-    """Create our own deprecation class, since Python >= 2.7
-    silences deprecations by default.
 
+def _get_stack_rank(func):
+    """Return function rank in the call stack."""
+    if _is_wrapped(func):
+        return 1 + _get_stack_rank(func.__wrapped__)
+    else:
+        return 0
+
+
+def _is_wrapped(func):
+    return "__wrapped__" in dir(func)
+
+
+def _get_stack_length(func):
+    """Return function call stack length."""
+    return _get_stack_rank(func.__globals__.get(func.__name__, func))
+
+
+class _DecoratorBaseClass:
+    """Used to manage decorators' warnings stacklevel.
+
+    The `_stack_length` class variable is used to store the number of
+    times a function is wrapped by a decorator.
+
+    Let `stack_length` be the total number of times a decorated
+    function is wrapped, and `stack_rank` be the rank of the decorator
+    in the decorators stack. The stacklevel of a warning is then
+    `stacklevel = 1 + stack_length - stack_rank`.
     """
 
-    pass
+    _stack_length = {}
+
+    def get_stack_length(self, func):
+        return self._stack_length.get(func.__name__,
+                                      _get_stack_length(func))
 
 
-class change_default_value:
+class change_default_value(_DecoratorBaseClass):
     """Decorator for changing the default value of an argument.
 
     Parameters
@@ -48,6 +79,7 @@ def __call__(self, func):
         arg_idx = list(parameters.keys()).index(self.arg_name)
         old_value = parameters[self.arg_name].default
 
+        stack_rank = _get_stack_rank(func)
         if self.warning_msg is None:
             self.warning_msg = (
                 f"The new recommended value for {self.arg_name} is "
@@ -59,15 +91,17 @@ def __call__(self, func):
 
         @functools.wraps(func)
         def fixed_func(*args, **kwargs):
+            stacklevel = 1 + self.get_stack_length(func) - stack_rank
             if len(args) < arg_idx + 1 and self.arg_name not in kwargs.keys():
                 # warn that arg_name default value changed:
-                warnings.warn(self.warning_msg, FutureWarning, stacklevel=2)
+                warnings.warn(self.warning_msg, FutureWarning,
+                              stacklevel=stacklevel)
             return func(*args, **kwargs)
 
         return fixed_func
 
 
-class remove_arg:
+class remove_arg(_DecoratorBaseClass):
     """Decorator to remove an argument from function's signature.
 
     Parameters
@@ -100,21 +134,25 @@ def __call__(self, func):
         if self.help_msg is not None:
             warning_msg += f" {self.help_msg}"
 
+        stack_rank = _get_stack_rank(func)
+
         @functools.wraps(func)
         def fixed_func(*args, **kwargs):
+            stacklevel = 1 + self.get_stack_length(func) - stack_rank
             if len(args) > arg_idx or self.arg_name in kwargs.keys():
                 # warn that arg_name is deprecated
-                warnings.warn(warning_msg, FutureWarning, stacklevel=2)
+                warnings.warn(warning_msg, FutureWarning,
+                              stacklevel=stacklevel)
             return func(*args, **kwargs)
 
         return fixed_func
 
 
-def docstring_add_deprecated(func, kwarg_mapping, deprecated_version):
+def _docstring_add_deprecated(func, kwarg_mapping, deprecated_version):
     """Add deprecated kwarg(s) to the "Other Params" section of a docstring.
 
     Parameters
-    ---------
+    ----------
     func : function
         The function whose docstring we wish to update.
     kwarg_mapping : dict
@@ -176,7 +214,7 @@ def docstring_add_deprecated(func, kwarg_mapping, deprecated_version):
     return final_docstring
 
 
-class deprecate_kwarg:
+class deprecate_kwarg(_DecoratorBaseClass):
     """Decorator ensuring backward compatibility when argument names are
     modified in a function definition.
 
@@ -213,14 +251,18 @@ def __init__(self, kwarg_mapping, deprecated_version, warning_msg=None,
 
     def __call__(self, func):
 
+        stack_rank = _get_stack_rank(func)
+
         @functools.wraps(func)
         def fixed_func(*args, **kwargs):
+            stacklevel = 1 + self.get_stack_length(func) - stack_rank
             for old_arg, new_arg in self.kwarg_mapping.items():
                 if old_arg in kwargs:
                     #  warn that the function interface has changed:
                     warnings.warn(self.warning_msg.format(
                         old_arg=old_arg, func_name=func.__name__,
-                        new_arg=new_arg), FutureWarning, stacklevel=2)
+                        new_arg=new_arg), FutureWarning,
+                        stacklevel=stacklevel)
                     # Substitute new_arg to old_arg
                     kwargs[new_arg] = kwargs.pop(old_arg)
 
@@ -228,13 +270,13 @@ def fixed_func(*args, **kwargs):
             return func(*args, **kwargs)
 
         if func.__doc__ is not None:
-            newdoc = docstring_add_deprecated(func, self.kwarg_mapping,
-                                              self.deprecated_version)
+            newdoc = _docstring_add_deprecated(func, self.kwarg_mapping,
+                                               self.deprecated_version)
             fixed_func.__doc__ = newdoc
         return fixed_func
 
 
-class channel_as_last_axis():
+class channel_as_last_axis:
     """Decorator for automatically making channels axis last for all arrays.
 
     This decorator reorders axes for compatibility with functions that only
@@ -311,53 +353,66 @@ def fixed_func(*args, **kwargs):
         return fixed_func
 
 
-class deprecated(object):
-    """Decorator to mark deprecated functions with warning.
+class deprecate_func(_DecoratorBaseClass):
+    """Decorate a deprecated function and warn when it is called.
 
     Adapted from <http://wiki.python.org/moin/PythonDecoratorLibrary>.
 
     Parameters
     ----------
-    alt_func : str
-        If given, tell user what function to use instead.
-    behavior : {'warn', 'raise'}
-        Behavior during call to deprecated function: 'warn' = warn user that
-        function is deprecated; 'raise' = raise error.
+    deprecated_version : str
+        The package version when the deprecation was introduced.
     removed_version : str
         The package version in which the deprecated function will be removed.
+    hint : str, optional
+        A hint on how to address this deprecation,
+        e.g., "Use `skimage.submodule.alternative_func` instead."
+    Examples
+    --------
+    >>> @deprecate_func(
+    ...     deprecated_version="1.0.0",
+    ...     removed_version="1.2.0",
+    ...     hint="Use `bar` instead."
+    ... )
+    ... def foo():
+    ...     pass
+    Calling ``foo`` will warn with::
+        FutureWarning: `foo` is deprecated since version 1.0.0
+        and will be removed in version 1.2.0. Use `bar` instead.
     """
 
-    def __init__(self, alt_func=None, behavior='warn', removed_version=None):
-        self.alt_func = alt_func
-        self.behavior = behavior
+    def __init__(self, *, deprecated_version, removed_version=None, hint=None):
+        self.deprecated_version = deprecated_version
         self.removed_version = removed_version
+        self.hint = hint
 
     def __call__(self, func):
 
-        alt_msg = ''
-        if self.alt_func is not None:
-            alt_msg = f' Use ``{self.alt_func}`` instead.'
-        rmv_msg = ''
-        if self.removed_version is not None:
-            rmv_msg = f' and will be removed in version {self.removed_version}'
+        message = (
+            f"`{func.__name__}` is deprecated since version "
+            f"{self.deprecated_version}"
+        )
+        if self.removed_version:
+            message += (
+                f" and will be removed in version {self.removed_version}."
+            )
+        if self.hint:
+            message += f" {self.hint.rstrip('.')}."
 
-        msg = f'Function ``{func.__name__}`` is deprecated{rmv_msg}.{alt_msg}'
+        stack_rank = _get_stack_rank(func)
 
         @functools.wraps(func)
         def wrapped(*args, **kwargs):
-            if self.behavior == 'warn':
-                func_code = func.__code__
-                warnings.simplefilter('always', skimage_deprecation)
-                warnings.warn_explicit(msg,
-                                       category=skimage_deprecation,
-                                       filename=func_code.co_filename,
-                                       lineno=func_code.co_firstlineno + 1)
-            elif self.behavior == 'raise':
-                raise skimage_deprecation(msg)
+            stacklevel = 1 + self.get_stack_length(func) - stack_rank
+            warnings.warn(
+                message,
+                category=FutureWarning,
+                stacklevel=stacklevel
+            )
             return func(*args, **kwargs)
 
         # modify doc string to display deprecation warning
-        doc = '**Deprecated function**.' + alt_msg
+        doc = f'**Deprecated:** {message}'
         if wrapped.__doc__ is None:
             wrapped.__doc__ = doc
         else:
diff --git a/python/cucim/src/cucim/skimage/color/__init__.py b/python/cucim/src/cucim/skimage/color/__init__.py
index 390668712..ef9049f34 100644
--- a/python/cucim/src/cucim/skimage/color/__init__.py
+++ b/python/cucim/src/cucim/skimage/color/__init__.py
@@ -9,12 +9,14 @@
                         rgb_from_bex, rgb_from_bpx, rgb_from_bro, rgb_from_fgx,
                         rgb_from_gdx, rgb_from_hax, rgb_from_hdx, rgb_from_hed,
                         rgb_from_hpx, rgb_from_rbd, rgba2rgb, rgbcie2rgb,
-                        separate_stains, xyz2lab, xyz2luv, xyz2rgb, ycbcr2rgb,
-                        ydbdr2rgb, yiq2rgb, ypbpr2rgb, yuv2rgb)
+                        separate_stains, xyz2lab, xyz2luv, xyz2rgb,
+                        xyz_tristimulus_values, ycbcr2rgb, ydbdr2rgb, yiq2rgb,
+                        ypbpr2rgb, yuv2rgb)
 from .colorlabel import color_dict, label2rgb
 from .delta_e import deltaE_cie76, deltaE_ciede94, deltaE_ciede2000, deltaE_cmc
 
 __all__ = ['convert_colorspace',
+           'xyz_tristimulus_values',
            'rgba2rgb',
            'rgb2hsv',
            'hsv2rgb',
diff --git a/python/cucim/src/cucim/skimage/color/colorconv.py b/python/cucim/src/cucim/skimage/color/colorconv.py
index ece630492..f7a1e3678 100644
--- a/python/cucim/src/cucim/skimage/color/colorconv.py
+++ b/python/cucim/src/cucim/skimage/color/colorconv.py
@@ -56,7 +56,7 @@
 from scipy import linalg
 
 from .._shared.utils import (_supported_float_type, channel_as_last_axis,
-                             identity)
+                             deprecate_func, identity)
 from ..util import dtype, dtype_limits
 
 
@@ -546,7 +546,7 @@ def hsv2rgb(hsv, *, channel_axis=-1):
 #    ----------
 #    .. [1] https://en.wikipedia.org/wiki/Standard_illuminant
 
-illuminants = \
+_illuminants = \
     {"A": {'2': (1.098466069456375, 1, 0.3558228003436005),
            '10': (1.111420406956693, 1, 0.3519978321919493),
            'R': (1.098466069456375, 1, 0.3558228003436005)},
@@ -573,7 +573,83 @@ def hsv2rgb(hsv, *, channel_axis=-1):
            'R': (1.0, 1.0, 1.0)}}
 
 
-def get_xyz_coords(illuminant, observer):
+def xyz_tristimulus_values(*, illuminant, observer, dtype=None):
+    """Get the CIE XYZ tristimulus values.
+
+    Given an illuminant and observer, this function returns the CIE XYZ
+    tristimulus values [2]_ scaled such that :math:`Y = 1`.
+
+    Parameters
+    ----------
+    illuminant : {"A", "B", "C", "D50", "D55", "D65", "D75", "E"}
+        The name of the illuminant (the function is NOT case sensitive).
+    observer : {"2", "10", "R"}
+        One of: 2-degree observer, 10-degree observer, or 'R' observer as in
+        R function ``grDevices::convertColor`` [3]_.
+    dtype : np.dtype, optional
+        This argument is ignored in the cuCIM implementation of
+        `xyz_tristimulus_values` since an array is not returned. The output is
+        always a 3-tuple of float.
+
+    Returns
+    -------
+    values : 3-tuple of float
+        Three elements :math:`X, Y, Z` containing the CIE XYZ tristimulus values
+        of the given illuminant.
+
+    Raises
+    ------
+    ValueError
+        If either the illuminant or the observer angle are not supported or
+        unknown.
+
+    References
+    ----------
+    .. [1] https://en.wikipedia.org/wiki/Standard_illuminant#White_points_of_standard_illuminants
+    .. [2] https://en.wikipedia.org/wiki/CIE_1931_color_space#Meaning_of_X,_Y_and_Z
+    .. [3] https://www.rdocumentation.org/packages/grDevices/versions/3.6.2/topics/convertColor
+
+    Notes
+    -----
+    The return type of this function differs from the one in scikit-image as it
+    always returns a 3-tuple of float rather than an array with a
+    user-specified dtype.
+
+    The CIE XYZ tristimulus values are calculated from :math:`x, y` [1]_, using the
+    formula
+
+    .. math:: X = x / y
+
+    .. math:: Y = 1
+
+    .. math:: Z = (1 - x - y) / y
+
+    The only exception is the illuminant "D65" with aperture angle 2° for
+    backward-compatibility reasons.
+
+    Examples
+    --------
+    Get the CIE XYZ tristimulus values for a "D65" illuminant for a 10 degree
+    field of view
+
+    >>> xyz_tristimulus_values(illuminant="D65", observer="10")
+    array([0.94809668, 1.        , 1.07305136])
+    """  # noqa
+    illuminant = illuminant.upper()
+    observer = observer.upper()
+    try:
+        return _illuminants[illuminant][observer]
+    except KeyError:
+        raise ValueError(f'Unknown illuminant/observer combination '
+                         f'(`{illuminant}`, `{observer}`)')
+
+
+@deprecate_func(
+    hint="Use `skimage.color.xyz_tristimulus_values` instead.",
+    deprecated_version="23.08",
+    removed_version="24.06",
+)
+def get_xyz_coords(illuminant, observer, dtype=float):
     """Get the XYZ coordinates of the given illuminant and observer [1]_.
 
     Parameters
@@ -602,13 +678,7 @@ def get_xyz_coords(illuminant, observer):
     ----------
     .. [1] https://en.wikipedia.org/wiki/Standard_illuminant
     """
-    illuminant = illuminant.upper()
-    observer = observer.upper()
-    try:
-        return illuminants[illuminant][observer]
-    except KeyError:
-        raise ValueError(f'Unknown illuminant/observer combination '
-                         f'(`{illuminant}`, `{observer}`)')
+    return xyz_tristimulus_values(illuminant=illuminant, observer=observer)
 
 
 # Haematoxylin-Eosin-DAB colorspace
@@ -625,7 +695,6 @@ def get_xyz_coords(illuminant, observer):
                          [0.27, 0.57, 0.78]])
 hed_from_rgb = linalg.inv(rgb_from_hed)
 
-
 # Following matrices are adapted form the Java code written by G.Landini.
 # The original code is available at:
 # https://web.archive.org/web/20160624145052/http://www.mecourse.com/landinig/software/cdeconv/cdeconv.html
@@ -1071,9 +1140,12 @@ def gray2rgb(image, *, channel_axis=-1):
 @cp.memoize(for_each_device=True)
 def _get_xyz_to_lab_kernel(xyz_ref_white, name='xyz2lab'):
     _xyz_to_lab = f"""
+        // scale by CIE XYZ tristimulus values of the reference white point
         arr[3*i] /= {xyz_ref_white[0]};
         arr[3*i + 1] /= {xyz_ref_white[1]};
         arr[3*i + 2] /= {xyz_ref_white[2]};
+
+        // Nonlinear distortion and linear transformation
         for (int ch=0; ch < 3; ch++)
         {{
             if (arr[3*i + ch] > 0.008856)
@@ -1083,6 +1155,8 @@ def _get_xyz_to_lab_kernel(xyz_ref_white, name='xyz2lab'):
                 arr[3*i + ch] = 7.787 * arr[3*i + ch] + 16.0 / 116.0;
             }}
         }}
+
+        // Vector scaling
         lab[3*i] = (116. * arr[3*i + 1]) - 16.0;
         lab[3*i + 1] = 500.0 * (arr[3*i] - arr[3*i + 1]);
         lab[3*i + 2] = 200.0 * (arr[3*i + 1] - arr[3*i + 2]);
@@ -1130,13 +1204,13 @@ def xyz2lab(xyz, illuminant="D65", observer="2", *, channel_axis=-1):
     Notes
     -----
     By default Observer="2", Illuminant="D65". CIE XYZ tristimulus values
-    x_ref=95.047, y_ref=100., z_ref=108.883. See function `get_xyz_coords` for
-    a list of supported illuminants.
+    x_ref=95.047, y_ref=100., z_ref=108.883. See function
+    :func:`~.xyz_tristimulus_values` for a list of supported illuminants.
 
     References
     ----------
-    .. [1] http://www.easyrgb.com/index.php?X=MATH&H=07
-    .. [2] https://en.wikipedia.org/wiki/Lab_color_space
+    .. [1] http://www.easyrgb.com/en/math.php
+    .. [2] https://en.wikipedia.org/wiki/CIELAB_color_space
 
     Examples
     --------
@@ -1150,8 +1224,9 @@ def xyz2lab(xyz, illuminant="D65", observer="2", *, channel_axis=-1):
     xyz = _prepare_colorarray(xyz, force_copy=True, force_c_contiguous=True,
                               channel_axis=channel_axis)
 
-    xyz_ref_white = get_xyz_coords(illuminant, observer)
-
+    xyz_ref_white = xyz_tristimulus_values(
+        illuminant=illuminant, observer=observer
+    )
     name = f'xyz2lab_{xyz.dtype.char}'
     kern = _get_xyz_to_lab_kernel(xyz_ref_white, name=name)
     lab = cp.empty_like(xyz)
@@ -1222,7 +1297,7 @@ def lab2xyz(lab, illuminant="D65", observer="2", *, channel_axis=-1):
     Returns
     -------
     out : (..., 3, ...) ndarray
-        The image in XYZ color space. Same dimensions as input.
+        The image in XYZ color space, of same shape as input.
 
     Raises
     ------
@@ -1237,18 +1312,49 @@ def lab2xyz(lab, illuminant="D65", observer="2", *, channel_axis=-1):
     Notes
     -----
     The CIE XYZ tristimulus values are x_ref = 95.047, y_ref = 100., and
-    z_ref = 108.883. See function :func:`~.get_xyz_coords` for a list of
+    z_ref = 108.883. See function :func:`~.xyz_tristimulus_values` for a list of
     supported illuminants.
 
+    See Also
+    --------
+    xyz2lab
+
     References
     ----------
-    .. [1] http://www.easyrgb.com/index.php?X=MATH&H=07
+    .. [1] http://www.easyrgb.com/en/math.php
     .. [2] https://en.wikipedia.org/wiki/CIELAB_color_space
     """
+    xyz, n_invalid = _lab2xyz(lab, illuminant, observer, channel_axis)
+    if n_invalid > 0:
+        warn(
+            "Conversion from CIE-LAB to XYZ color space resulted in "
+            f"{n_invalid} negative Z values that have been clipped to zero",
+            stacklevel=3,
+        )
+
+    return xyz
+
+
+def _lab2xyz(lab, illuminant, observer, channel_axis):
+    """Convert CIE-LAB to XYZ color space.
+
+    Internal function for :func:`~.lab2xyz` and others. In addition to the
+    converted image, return the number of invalid pixels in the Z channel for
+    correct warning propagation.
+
+    Returns
+    -------
+    out : (..., 3, ...) ndarray
+        The image in XYZ format. Same dimensions as input.
+    n_invalid : int
+        Number of invalid pixels in the Z channel after conversion.
+    """
     lab = _prepare_colorarray(lab, force_c_contiguous=True,
                               channel_axis=channel_axis)
 
-    xyz_ref_white = get_xyz_coords(illuminant, observer)
+    xyz_ref_white = xyz_tristimulus_values(
+        illuminant=illuminant, observer=observer
+    )
 
     name = f'lab2xyz_{lab.dtype.char}'
     kern = _get_lab_to_xyz_kernel(xyz_ref_white, name=name)
@@ -1258,12 +1364,8 @@ def lab2xyz(lab, illuminant="D65", observer="2", *, channel_axis=-1):
     #       operations?
     warnings = cp.zeros(lab.shape[:-1], dtype=np.int32)
     kern(lab, xyz, warnings, size=lab.size // 3)
-
-    nwarn = int(cp.count_nonzero(warnings))
-    if nwarn > 0:  # synchronize!
-        warn(f'Color data out of range: Z < 0 in {nwarn} pixels',
-             stacklevel=3)
-    return xyz
+    n_invalid = int(cp.count_nonzero(warnings))  # synchronize!
+    return xyz, n_invalid
 
 
 @channel_as_last_axis()
@@ -1302,8 +1404,8 @@ def rgb2lab(rgb, illuminant="D65", observer="2", *, channel_axis=-1):
 
     This function uses rgb2xyz and xyz2lab.
     By default Observer="2", Illuminant="D65". CIE XYZ tristimulus values
-    x_ref=95.047, y_ref=100., z_ref=108.883. See function `get_xyz_coords` for
-    a list of supported illuminants.
+    x_ref=95.047, y_ref=100., z_ref=108.883. See function
+    :func:`~.xyz_tristimulus_values` for a list of supported illuminants.
 
     References
     ----------
@@ -1335,7 +1437,7 @@ def lab2rgb(lab, illuminant="D65", observer="2", *, channel_axis=-1):
     Returns
     -------
     out : (..., 3, ...) ndarray
-        The image in RGB format. Same dimensions as input.
+        The image in sRGB color space, of same shape as input.
 
     Raises
     ------
@@ -1344,17 +1446,28 @@ def lab2rgb(lab, illuminant="D65", observer="2", *, channel_axis=-1):
 
     Notes
     -----
-    This function uses lab2xyz and xyz2rgb.
+    This function uses :func:`~.lab2xyz` and :func:`~.xyz2rgb`.
     The CIE XYZ tristimulus values are x_ref = 95.047, y_ref = 100., and
-    z_ref = 108.883. See function :func:`~.get_xyz_coords` for a list of
+    z_ref = 108.883. See function :func:`~.xyz_tristimulus_values` for a list of
     supported illuminants.
 
+    See Also
+    --------
+    rgb2lab
+
     References
     ----------
     .. [1] https://en.wikipedia.org/wiki/Standard_illuminant
     .. [2] https://en.wikipedia.org/wiki/CIELAB_color_space
     """
-    return xyz2rgb(lab2xyz(lab, illuminant, observer))
+    xyz, n_invalid = _lab2xyz(lab, illuminant, observer, channel_axis)
+    if n_invalid != 0:
+        warn(
+            "Conversion from CIE-LAB, via XYZ to sRGB color space resulted in "
+            f"{n_invalid} negative Z values that have been clipped to zero",
+            stacklevel=3,
+        )
+    return xyz2rgb(xyz, channel_axis=channel_axis)
 
 
 @cp.memoize(for_each_device=True)
@@ -1442,12 +1555,12 @@ def xyz2luv(xyz, illuminant="D65", observer="2", *, channel_axis=-1):
     -----
     By default XYZ conversion weights use observer=2A. Reference whitepoint
     for D65 Illuminant, with XYZ tristimulus values of ``(95.047, 100.,
-    108.883)``. See function 'get_xyz_coords' for a list of supported
-    illuminants.
+    108.883)``. See function :func:`~.xyz_tristimulus_values` for a list of
+    supported illuminants.
 
     References
     ----------
-    .. [1] http://www.easyrgb.com/index.php?X=MATH&H=16#text16
+    .. [1] http://www.easyrgb.com/en/math.php
     .. [2] https://en.wikipedia.org/wiki/CIELUV
 
     Examples
@@ -1466,7 +1579,9 @@ def xyz2luv(xyz, illuminant="D65", observer="2", *, channel_axis=-1):
     xyz = _prepare_colorarray(xyz, force_c_contiguous=True,
                               channel_axis=channel_axis)
 
-    xyz_ref_white = get_xyz_coords(illuminant, observer)
+    xyz_ref_white = xyz_tristimulus_values(
+        illuminant=illuminant, observer=observer
+    )
     kern = _get_xyz_to_luv_kernel(xyz_ref_white, xyz.dtype)
     luv = cp.empty_like(xyz)
     kern(xyz, luv, size=xyz.size // 3)
@@ -1546,16 +1661,19 @@ def luv2xyz(luv, illuminant="D65", observer="2", *, channel_axis=-1):
     -----
     XYZ conversion weights use observer=2A. Reference whitepoint for D65
     Illuminant, with XYZ tristimulus values of ``(95.047, 100., 108.883)``. See
-    function 'get_xyz_coords' for a list of supported illuminants.
+    function :func:`~.xyz_tristimulus_values` for a list of supported
+    illuminants.
 
     References
     ----------
-    .. [1] http://www.easyrgb.com/index.php?X=MATH&H=16#text16
+    .. [1] http://www.easyrgb.com/en/math.php
     .. [2] https://en.wikipedia.org/wiki/CIELUV
     """
     luv = _prepare_colorarray(luv, force_c_contiguous=True,
                               channel_axis=channel_axis)
-    xyz_ref_white = get_xyz_coords(illuminant, observer)
+    xyz_ref_white = xyz_tristimulus_values(
+        illuminant=illuminant, observer=observer
+    )
     kern = _get_luv_to_xyz_kernel(xyz_ref_white, luv.dtype)
     xyz = cp.empty_like(luv)
     kern(luv, xyz, size=luv.size // 3)
@@ -1591,9 +1709,8 @@ def rgb2luv(rgb, *, channel_axis=-1):
 
     References
     ----------
-    .. [1] http://www.easyrgb.com/index.php?X=MATH&H=16#text16
-    .. [2] http://www.easyrgb.com/index.php?X=MATH&H=02#text2
-    .. [3] https://en.wikipedia.org/wiki/CIELUV
+    .. [1] http://www.easyrgb.com/en/math.php
+    .. [2] https://en.wikipedia.org/wiki/CIELUV
     """
     return xyz2luv(rgb2xyz(rgb))
 
diff --git a/python/cucim/src/cucim/skimage/color/delta_e.py b/python/cucim/src/cucim/skimage/color/delta_e.py
index 14ed971db..bf07162a8 100644
--- a/python/cucim/src/cucim/skimage/color/delta_e.py
+++ b/python/cucim/src/cucim/skimage/color/delta_e.py
@@ -29,7 +29,7 @@
 
 def _float_inputs(lab1, lab2, allow_float32=True):
     if allow_float32:
-        float_dtype = _supported_float_type([lab1.dtype, lab2.dtype])
+        float_dtype = _supported_float_type((lab1.dtype, lab2.dtype))
     else:
         float_dtype = cp.float64
     lab1 = lab1.astype(float_dtype, copy=False)
@@ -375,7 +375,7 @@ def get_dH2(lab1, lab2, *, channel_axis=-1):
     """
     # This function needs double precision internally for accuracy
     input_is_float_32 = _supported_float_type(
-        [lab1.dtype, lab2.dtype]
+        (lab1.dtype, lab2.dtype)
     ) == cp.float32
     lab1, lab2 = _float_inputs(lab1, lab2, allow_float32=False)
     a1, b1 = cp.moveaxis(lab1, source=channel_axis, destination=0)[1:3]
diff --git a/python/cucim/src/cucim/skimage/color/tests/test_colorconv.py b/python/cucim/src/cucim/skimage/color/tests/test_colorconv.py
index 0d50cf119..f3bd45a1e 100644
--- a/python/cucim/src/cucim/skimage/color/tests/test_colorconv.py
+++ b/python/cucim/src/cucim/skimage/color/tests/test_colorconv.py
@@ -666,9 +666,13 @@ def test_lab_full_gamut(self):
         a, b = cp.meshgrid(cp.arange(-100, 100), cp.arange(-100, 100))
         L = cp.ones(a.shape)
         lab = cp.dstack((L, a, b))
+        regex = (
+            "Conversion from CIE-LAB to XYZ color space resulted in "
+            "\\d+ negative Z values that have been clipped to zero"
+        )
         for value in [0, 10, 20]:
             lab[:, :, 0] = value
-            with expected_warnings(['Color data out of range']):
+            with pytest.warns(UserWarning, match=regex):
                 lab2xyz(lab)
 
     @pytest.mark.parametrize("channel_axis", [0, 1, -1, -2])
@@ -800,6 +804,17 @@ def test_rgb2yiq_conversion(self):
                         )
         assert_array_almost_equal(yiq, gt, decimal=2)
 
+    @pytest.mark.parametrize("func", [lab2rgb, lab2xyz])
+    def test_warning_stacklevel(self, func):
+        regex = (
+            "Conversion from CIE-LAB.* XYZ.*color space resulted in "
+            "1 negative Z values that have been clipped to zero"
+        )
+        with pytest.warns(UserWarning, match=regex) as messages:
+            func(lab=cp.array([[[0, 0, 300.]]]))
+        assert len(messages) == 1
+        assert messages[0].filename == __file__, "warning points at wrong file"
+
 
 def test_gray2rgb():
     x = cp.asarray([0, 0.5, 1])
diff --git a/python/cucim/src/cucim/skimage/exposure/_adapthist.py b/python/cucim/src/cucim/skimage/exposure/_adapthist.py
index 4d42e2ae3..939601973 100644
--- a/python/cucim/src/cucim/skimage/exposure/_adapthist.py
+++ b/python/cucim/src/cucim/skimage/exposure/_adapthist.py
@@ -4,14 +4,7 @@
 
 http://tog.acm.org/resources/GraphicsGems/
 
-The Graphics Gems code is copyright-protected.  In other words, you cannot
-claim the text of the code as your own and resell it. Using the code is
-permitted in any program, product, or library, non-commercial or commercial.
-Giving credit is not required, though is a nice gesture.  The code comes as-is,
-and if there are any flaws or problems with any Gems code, nobody involved with
-Gems - authors, editors, publishers, or webmasters - are to be held
-responsible.  Basically, don't be a jerk, and remember that anything free
-comes with no guarantee.
+Relicensed with permission of the author under the Modified BSD license.
 """
 import functools
 import itertools
@@ -97,7 +90,7 @@ def equalize_adapthist(image, kernel_size=None,
     elif isinstance(kernel_size, numbers.Number):
         kernel_size = (kernel_size,) * image.ndim
     elif len(kernel_size) != image.ndim:
-        ValueError(f'Incorrect value of `kernel_size`: {kernel_size}')
+        raise ValueError(f'Incorrect value of `kernel_size`: {kernel_size}')
 
     kernel_size = [int(k) for k in kernel_size]
 
diff --git a/python/cucim/src/cucim/skimage/exposure/exposure.py b/python/cucim/src/cucim/skimage/exposure/exposure.py
index 4230f36c9..c6038cae4 100644
--- a/python/cucim/src/cucim/skimage/exposure/exposure.py
+++ b/python/cucim/src/cucim/skimage/exposure/exposure.py
@@ -72,8 +72,14 @@ def _bincount_histogram(image, source_range, bin_centers=None):
     """
     if bin_centers is None:
         bin_centers = _bincount_histogram_centers(image, source_range)
-    image_min, image_max = bin_centers[0], bin_centers[-1]
-    image = _offset_array(image, image_min.item(), image_max.item())  # synchronize  # noqa
+    image_min, image_max = bin_centers[0].item(), bin_centers[-1].item()
+    image = _offset_array(image, image_min, image_max)  # synchronize  # noqa
+
+    # Casting back to unsigned dtype seems necessary to avoid incorrect
+    # results for larger integer ranges with CUDA 12.x.
+    unsigned_dtype_char = image.dtype.char.upper()
+    image = image.astype(unsigned_dtype_char, copy=False)
+
     hist = cp.bincount(
         image.ravel(), minlength=image_max - min(image_min, 0) + 1
     )
@@ -178,7 +184,9 @@ def _get_numpy_hist_range(image, source_range):
     elif source_range == 'dtype':
         hist_range = dtype_limits(image, clip_negative=False)
     else:
-        ValueError('Wrong value for the `source_range` argument')
+        raise ValueError(
+            f'Incorrect value for `source_range` argument: {source_range}'
+        )
     return hist_range
 
 
diff --git a/python/cucim/src/cucim/skimage/exposure/tests/test_exposure.py b/python/cucim/src/cucim/skimage/exposure/tests/test_exposure.py
index 7087611f0..11f6fb673 100644
--- a/python/cucim/src/cucim/skimage/exposure/tests/test_exposure.py
+++ b/python/cucim/src/cucim/skimage/exposure/tests/test_exposure.py
@@ -25,7 +25,8 @@
 
 def test_wrong_source_range():
     im = cp.array([-1, 100], dtype=cp.int8)
-    with pytest.raises(ValueError):
+    match = "Incorrect value for `source_range` argument"
+    with pytest.raises(ValueError, match=match):
         frequencies, bin_centers = exposure.histogram(
             im, source_range="foobar"
         )
@@ -592,6 +593,12 @@ def norm_brightness_err(img1, img2):
     return nbe
 
 
+def test_adapthist_incorrect_kernel_size():
+    img = cp.ones((8, 8), dtype=float)
+    with pytest.raises(ValueError, match="Incorrect value of `kernel_size`"):
+        exposure.equalize_adapthist(img, (3, 3, 3))
+
+
 # Test Gamma Correction
 # =====================
 
diff --git a/python/cucim/src/cucim/skimage/feature/__init__.py b/python/cucim/src/cucim/skimage/feature/__init__.py
index d26dfb347..fcfa62c03 100644
--- a/python/cucim/src/cucim/skimage/feature/__init__.py
+++ b/python/cucim/src/cucim/skimage/feature/__init__.py
@@ -1,4 +1,3 @@
-from .._shared.utils import deprecated
 from ._basic_features import multiscale_basic_features
 from ._canny import canny
 from ._daisy import daisy
diff --git a/python/cucim/src/cucim/skimage/feature/_canny.py b/python/cucim/src/cucim/skimage/feature/_canny.py
index edc826fab..1d60dbf86 100644
--- a/python/cucim/src/cucim/skimage/feature/_canny.py
+++ b/python/cucim/src/cucim/skimage/feature/_canny.py
@@ -358,6 +358,9 @@ def canny(image, sigma=1., low_threshold=None, high_threshold=None, mask=None,
     # mask by one and then mask the output. We also mask out the border points
     # because who knows what lies beyond the edge of the image?
 
+    if (image.dtype.kind in 'iu' and image.dtype.itemsize >= 8):
+        raise ValueError("64-bit or larger integer images are not supported")
+
     check_nD(image, 2)
     dtype_max = dtype_limits(image, clip_negative=False)[1]
 
diff --git a/python/cucim/src/cucim/skimage/feature/tests/test_canny.py b/python/cucim/src/cucim/skimage/feature/tests/test_canny.py
index 6e06114fc..8f3ab90ff 100644
--- a/python/cucim/src/cucim/skimage/feature/tests/test_canny.py
+++ b/python/cucim/src/cucim/skimage/feature/tests/test_canny.py
@@ -77,7 +77,7 @@ def test_mask_none(self):
         assert cp.all(result1 == result2)
 
     @cp.testing.with_requires("scikit-image>=0.18")
-    @pytest.mark.parametrize('image_dtype', [cp.uint8, cp.int64, cp.float32,
+    @pytest.mark.parametrize('image_dtype', [cp.uint8, cp.int32, cp.float32,
                                              cp.float64])
     def test_use_quantiles(self, image_dtype):
         dtype = cp.dtype(image_dtype)
@@ -146,3 +146,24 @@ def test_dtype(self):
             feature.canny(image_float, 1.0, low, high),
             feature.canny(image_uint8, 1.0, 255 * low, 255 * high)
         )
+
+    def test_full_mask_matches_no_mask(self):
+        """The masked and unmasked algorithms should return the same result.
+
+        """
+        image = cp.array(data.camera())
+
+        for mode in ('constant', 'nearest', 'reflect'):
+            cp.testing.assert_array_equal(
+                feature.canny(image, mode=mode),
+                feature.canny(image, mode=mode,
+                              mask=cp.ones_like(image, dtype=bool))
+            )
+
+    @pytest.mark.parametrize('dtype', (cp.int64, cp.uint64))
+    def test_unsupported_int64(self, dtype):
+        image = cp.zeros((10, 10), dtype=dtype)
+        image[3, 3] = cp.iinfo(dtype).max
+        match = "64-bit or larger integer images are not supported"
+        with pytest.raises(ValueError, match=match):
+            feature.canny(image)
diff --git a/python/cucim/src/cucim/skimage/filters/__init__.py b/python/cucim/src/cucim/skimage/filters/__init__.py
index 29ebabac1..32cfe78d8 100644
--- a/python/cucim/src/cucim/skimage/filters/__init__.py
+++ b/python/cucim/src/cucim/skimage/filters/__init__.py
@@ -4,7 +4,9 @@
     __name__,
     # submodules={'rank'},
     submod_attrs={
-        'lpi_filter': ['filter_inverse', 'wiener', 'LPIFilter2D'],
+        'lpi_filter': [
+            'filter_forward', 'filter_inverse', 'wiener', 'LPIFilter2D'
+        ],
         '_gaussian': ['gaussian', 'difference_of_gaussians'],
         'edges': ['sobel', 'sobel_h', 'sobel_v',
                   'scharr', 'scharr_h', 'scharr_v',
@@ -14,12 +16,14 @@
                   'farid', 'farid_h', 'farid_v'],
         '_rank_order': ['rank_order'],
         '_gabor': ['gabor_kernel', 'gabor'],
-        'thresholding': ['threshold_local', 'threshold_otsu', 'threshold_yen',
-                         'threshold_isodata', 'threshold_li', 'threshold_minimum',
-                         'threshold_mean', 'threshold_triangle',
-                         'threshold_niblack', 'threshold_sauvola',
-                         'threshold_multiotsu', 'try_all_threshold',
-                         'apply_hysteresis_threshold'],
+        'thresholding': [
+            'threshold_local', 'threshold_otsu', 'threshold_yen',
+            'threshold_isodata', 'threshold_li', 'threshold_minimum',
+            'threshold_mean', 'threshold_triangle',
+            'threshold_niblack', 'threshold_sauvola',
+            'threshold_multiotsu', 'try_all_threshold',
+            'apply_hysteresis_threshold'
+        ],
         'ridges': ['meijering', 'sato', 'frangi', 'hessian'],
         '_median': ['median'],
         '_sparse': ['correlate_sparse'],
diff --git a/python/cucim/src/cucim/skimage/filters/lpi_filter.py b/python/cucim/src/cucim/skimage/filters/lpi_filter.py
index d6cd59c0b..e443a0723 100644
--- a/python/cucim/src/cucim/skimage/filters/lpi_filter.py
+++ b/python/cucim/src/cucim/skimage/filters/lpi_filter.py
@@ -7,7 +7,7 @@
 import numpy as np
 from cupyx.scipy import fft
 
-from .._shared.utils import _supported_float_type, check_nD, deprecated
+from .._shared.utils import _supported_float_type, check_nD, deprecate_func
 
 eps = np.finfo(float).eps
 
@@ -19,7 +19,7 @@ def _min_limit(x, val=eps):
 
 def _center(x, oshape):
     """Return an array of shape ``oshape`` from the center of array ``x``."""
-    start = (np.array(x.shape) - np.array(oshape)) // 2 + 1
+    start = (np.array(x.shape) - np.array(oshape)) // 2
     out = x[tuple(slice(s, s + n) for s, n in zip(start, oshape))]
     return out
 
@@ -65,11 +65,10 @@ def __init__(self, impulse_response, **filter_params):
 
         Examples
         --------
-        Gaussian filter: Use a 1-D gaussian in each direction without
-        normalization coefficients.
+        Gaussian filter without normalization of coefficients:
 
         >>> def filt_func(r, c, sigma = 1):
-        ...     return cp.exp(-cp.hypot(r, c)/sigma)
+        ...     return cp.exp(-(r**2 + c**2)/(2 * sigma**2))
         >>> filter = LPIFilter2D(filt_func)
 
         """
@@ -83,14 +82,22 @@ def __init__(self, impulse_response, **filter_params):
     def _prepare(self, data):
         """Calculate filter and data FFT in preparation for filtering."""
         dshape = np.array(data.shape)
-        dshape += dshape % 2 == 0  # all filter dimensions must be uneven
-        oshape = np.array(data.shape) * 2 - 1
+        # all filter dimensions must be uneven
+        even_offset = tuple(int(s % 2 == 0) for s in data.shape)
+        dshape = tuple(
+            s + offset for s, offset in zip(data.shape, even_offset)
+        )
+
+        oshape = tuple(s * 2 - 1 for s in data.shape)
 
         float_dtype = _supported_float_type(data.dtype)
         data = data.astype(float_dtype, copy=False)
 
         if self._cache is None or np.any(self._cache.shape != oshape):
-            coords = cp.mgrid[[slice(0, float(n)) for n in dshape]]
+            coords = cp.mgrid[
+                [slice(0 + offset, float(n + offset))
+                 for (n, offset) in zip(dshape, even_offset)]
+            ]
             # this steps over two sets of coordinates,
             # not over the coordinates individually
             for k, coord in enumerate(coords):
@@ -127,7 +134,7 @@ def __call__(self, data):
         return out
 
 
-def filter_forward(data, impulse_response=None, filter_params={},
+def filter_forward(data, impulse_response=None, filter_params=None,
                    predefined_filter=None):
     """Apply the given filter to data.
 
@@ -137,7 +144,7 @@ def filter_forward(data, impulse_response=None, filter_params={},
         Input data.
     impulse_response : callable `f(r, c, **filter_params)`
         Impulse response of the filter.  See LPIFilter2D.__init__.
-    filter_params : dict
+    filter_params : dict, optional
         Additional keyword parameters to the impulse_response function.
 
     Other Parameters
@@ -148,32 +155,35 @@ def filter_forward(data, impulse_response=None, filter_params={},
 
     Examples
     --------
+    Gaussian filter without normalization:
 
-    Gaussian filter:
-
-    >>> import cupy as cp
-    >>> def filt_func(r, c):
-    ...     return cp.exp(-cp.hypot(r, c)/1)
+    >>> def filt_func(r, c, sigma=1):
+    ...     return cp.exp(-(r**2 + c**2)/(2 * sigma**2))
     >>>
     >>> from skimage import data
     >>> filtered = filter_forward(cp.array(data.coins()), filt_func)
 
     """
+    if filter_params is None:
+        filter_params = {}
     check_nD(data, 2, 'data')
     if predefined_filter is None:
         predefined_filter = LPIFilter2D(impulse_response, **filter_params)
     return predefined_filter(data)
 
 
-@deprecated(alt_func='cucim.skimage.filters.lpi_filter.filter_inverse',
-            removed_version='2023.06.00')
-def inverse(data, impulse_response=None, filter_params={}, max_gain=2,
+@deprecate_func(
+    hint="use cucim.skimage.filters.lpi_filter.filter_inverse instead",
+    deprecated_version="",
+    removed_version="2023.12.00",
+)
+def inverse(data, impulse_response=None, filter_params=None, max_gain=2,
             predefined_filter=None):
     return filter_inverse(data, impulse_response, filter_params,
                           max_gain, predefined_filter)
 
 
-def filter_inverse(data, impulse_response=None, filter_params={}, max_gain=2,
+def filter_inverse(data, impulse_response=None, filter_params=None, max_gain=2,
                    predefined_filter=None):
     """Apply the filter in reverse to the given data.
 
@@ -182,21 +192,24 @@ def filter_inverse(data, impulse_response=None, filter_params={}, max_gain=2,
     data : (M, N) ndarray
         Input data.
     impulse_response : callable `f(r, c, **filter_params)`
-        Impulse response of the filter.  See LPIFilter2D.__init__.
-    filter_params : dict
+        Impulse response of the filter.  See :class:`~.LPIFilter2D`. This is a
+        required argument unless a `predifined_filter` is provided.
+    filter_params : dict, optional
         Additional keyword parameters to the impulse_response function.
-    max_gain : float
+    max_gain : float, optional
         Limit the filter gain.  Often, the filter contains zeros, which would
         cause the inverse filter to have infinite gain.  High gain causes
         amplification of artefacts, so a conservative limit is recommended.
 
     Other Parameters
     ----------------
-    predefined_filter : LPIFilter2D
+    predefined_filter : LPIFilter2D, optional
         If you need to apply the same filter multiple times over different
         images, construct the LPIFilter2D and specify it here.
 
     """
+    if filter_params is None:
+        filter_params = {}
     check_nD(data, 2, 'data')
     if predefined_filter is None:
         filt = LPIFilter2D(impulse_response, **filter_params)
@@ -213,7 +226,7 @@ def filter_inverse(data, impulse_response=None, filter_params={}, max_gain=2,
     return _center(cp.abs(fft.ifftshift(fft.ifftn(G * F))), data.shape)
 
 
-def wiener(data, impulse_response=None, filter_params={}, K=0.25,
+def wiener(data, impulse_response=None, filter_params=None, K=0.25,
            predefined_filter=None):
     """Minimum Mean Square Error (Wiener) inverse filter.
 
@@ -226,7 +239,7 @@ def wiener(data, impulse_response=None, filter_params={}, K=0.25,
         image.
     impulse_response : callable `f(r, c, **filter_params)`
         Impulse response of the filter.  See LPIFilter2D.__init__.
-    filter_params : dict
+    filter_params : dict, optional
         Additional keyword parameters to the impulse_response function.
 
     Other Parameters
@@ -236,6 +249,8 @@ def wiener(data, impulse_response=None, filter_params={}, K=0.25,
         images, construct the LPIFilter2D and specify it here.
 
     """
+    if filter_params is None:
+        filter_params = {}
     check_nD(data, 2, 'data')
 
     if not isinstance(K, float):
diff --git a/python/cucim/src/cucim/skimage/filters/ridges.py b/python/cucim/src/cucim/skimage/filters/ridges.py
index 5d861b8f7..3615278bc 100644
--- a/python/cucim/src/cucim/skimage/filters/ridges.py
+++ b/python/cucim/src/cucim/skimage/filters/ridges.py
@@ -14,13 +14,13 @@ class of ridge filters relies on the eigenvalues of the Hessian matrix of
 import cupy as cp
 import numpy as np
 
-from .._shared.utils import _supported_float_type, check_nD, deprecated
+from .._shared.utils import _supported_float_type, check_nD, deprecate_func
 from ..feature.corner import (_symmetric_compute_eigenvalues, hessian_matrix,
                               hessian_matrix_eigvals)
 from ..util import img_as_float
 
 
-@deprecated(removed_version="2023.06.01")
+@deprecate_func(deprecated_version="", removed_version="2023.06.01")
 def compute_hessian_eigenvalues(image, sigma, sorting='none',
                                 mode='constant', cval=0,
                                 use_gaussian_derivatives=False):
diff --git a/python/cucim/src/cucim/skimage/filters/thresholding.py b/python/cucim/src/cucim/skimage/filters/thresholding.py
index 367eb35d8..c2dedb2fa 100644
--- a/python/cucim/src/cucim/skimage/filters/thresholding.py
+++ b/python/cucim/src/cucim/skimage/filters/thresholding.py
@@ -704,6 +704,10 @@ def threshold_li(image, *, tolerance=None, initial_guess=None,
                                    weights=hist[foreground])
             mean_back = np.average(bin_centers[background],
                                    weights=hist[background])
+
+            if mean_back == 0:
+                break
+
             eps = 100 * np.finfo(float).eps
             mean_back = float(mean_back)
             mean_fore = float(mean_fore)
@@ -726,6 +730,9 @@ def threshold_li(image, *, tolerance=None, initial_guess=None,
             mean_fore = float(cp.mean(image[foreground]))
             mean_back = float(cp.mean(image[~foreground]))
 
+            if mean_back == 0:
+                break
+
             t_next = ((mean_back - mean_fore) /
                       (math.log(mean_back + eps) - math.log(mean_fore + eps)))
 
diff --git a/python/cucim/src/cucim/skimage/morphology/_skeletonize.py b/python/cucim/src/cucim/skimage/morphology/_skeletonize.py
index d4ca12fde..a58ba8681 100644
--- a/python/cucim/src/cucim/skimage/morphology/_skeletonize.py
+++ b/python/cucim/src/cucim/skimage/morphology/_skeletonize.py
@@ -153,10 +153,10 @@ def thin(image, max_num_iter=None):
 # --------- Skeletonization by medial axis transform --------
 
 
-def _get_tiebreaker(n, random_seed):
+def _get_tiebreaker(n, seed):
     # CuPy generator doesn't currently have the permutation method, so
     # fall back to cp.random.permutation instead.
-    cp.random.seed(random_seed)
+    cp.random.seed(seed)
     if n < 2 << 31:
         dtype = np.int32
     else:
@@ -165,7 +165,12 @@ def _get_tiebreaker(n, random_seed):
     return tiebreaker
 
 
-def medial_axis(image, mask=None, return_distance=False, *, random_state=None):
+@deprecate_kwarg(
+    {'random_state': 'seed'},
+    deprecated_version='23.08',
+    removed_version='24.06'
+)
+def medial_axis(image, mask=None, return_distance=False, *, seed=None):
     """Compute the medial axis transform of a binary image.
 
     Parameters
@@ -177,13 +182,12 @@ def medial_axis(image, mask=None, return_distance=False, *, random_state=None):
         value in `mask` are used for computing the medial axis.
     return_distance : bool, optional
         If true, the distance transform is returned as well as the skeleton.
-    random_state : {None, int, `numpy.random.Generator`}, optional
-        If `random_state` is None the `numpy.random.Generator` singleton is
+    seed : {None, int, `numpy.random.Generator`}, optional
+        If `seed` is None, the `numpy.random.Generator` singleton is used.
+        If `seed` is an int, a new ``Generator`` instance is used, seeded with
+        `seed`.
+        If `seed` is already a ``Generator`` instance, then that instance is
         used.
-        If `random_state` is an int, a new ``Generator`` instance is used,
-        seeded with `random_state`.
-        If `random_state` is already a ``Generator`` instance then that
-        instance is used.
 
         .. versionadded:: 0.19
 
@@ -295,7 +299,7 @@ def medial_axis(image, mask=None, return_distance=False, *, random_state=None):
     # We use a random # for tiebreaking. Assign each pixel in the image a
     # predictable, random # so that masking doesn't affect arbitrary choices
     # of skeletons
-    tiebreaker = _get_tiebreaker(n=distance.size, random_seed=random_state)
+    tiebreaker = _get_tiebreaker(n=distance.size, seed=seed)
     order = cp.lexsort(
         cp.stack(
             (tiebreaker, corner_score[masked_image], distance),
diff --git a/python/cucim/src/cucim/skimage/restoration/__init__.py b/python/cucim/src/cucim/skimage/restoration/__init__.py
index aa0b63ecb..07eaeb6b1 100644
--- a/python/cucim/src/cucim/skimage/restoration/__init__.py
+++ b/python/cucim/src/cucim/skimage/restoration/__init__.py
@@ -1,6 +1,6 @@
 from ._denoise import denoise_tv_chambolle
 from .deconvolution import richardson_lucy, unsupervised_wiener, wiener
-from .j_invariant import calibrate_denoiser
+from .j_invariant import calibrate_denoiser, denoise_invariant
 
 __all__ = [
     "wiener",
@@ -8,4 +8,5 @@
     "richardson_lucy",
     "denoise_tv_chambolle",
     "calibrate_denoiser",
+    "denoise_invariant",
 ]
diff --git a/python/cucim/src/cucim/skimage/restoration/deconvolution.py b/python/cucim/src/cucim/skimage/restoration/deconvolution.py
index 923505a4a..f9dda2e30 100644
--- a/python/cucim/src/cucim/skimage/restoration/deconvolution.py
+++ b/python/cucim/src/cucim/skimage/restoration/deconvolution.py
@@ -146,8 +146,10 @@ def wiener(image, psf, balance, reg=None, is_real=True, clip=True):
     return deconv
 
 
+@deprecate_kwarg({'random_state': 'seed'}, removed_version="23.08.00",
+                 deprecated_version="24.06.00")
 def unsupervised_wiener(image, psf, reg=None, user_params=None, is_real=True,
-                        clip=True, *, random_state=None):
+                        clip=True, *, seed=None):
     """Unsupervised Wiener-Hunt deconvolution.
 
     Return the deconvolution with a Wiener-Hunt approach, where the
@@ -172,13 +174,12 @@ def unsupervised_wiener(image, psf, reg=None, user_params=None, is_real=True,
     clip : boolean, optional
        True by default. If true, pixel values of the result above 1 or
        under -1 are thresholded for skimage pipeline compatibility.
-    random_state : {None, int, `cupy.random.Generator`}, optional
-        If `random_state` is None the `cupy.random.Generator` singleton is
+    seed : {None, int, `numpy.random.Generator`}, optional
+        If `seed` is None, the `numpy.random.Generator` singleton is used.
+        If `seed` is an int, a new ``Generator`` instance is used, seeded with
+        `seed`.
+        If `seed` is already a ``Generator`` instance, then that instance is
         used.
-        If `random_state` is an int, a new ``Generator`` instance is used,
-        seeded with `random_state`.
-        If `random_state` is already a ``Generator`` instance then that
-        instance is used.
 
     Returns
     -------
@@ -304,10 +305,10 @@ def unsupervised_wiener(image, psf, reg=None, user_params=None, is_real=True,
         data_spectrum = uft.ufft2(image)
 
     try:
-        rng = cp.random.default_rng(random_state)
+        rng = cp.random.default_rng(seed)
     except AttributeError:
         # older CuPy without default_rng
-        rng = cp.random.RandomState(random_state)
+        rng = cp.random.RandomState(seed)
 
     # Gibbs sampling
     for iteration in range(params["max_num_iter"]):
@@ -383,8 +384,6 @@ def unsupervised_wiener(image, psf, reg=None, user_params=None, is_real=True,
     return (x_postmean, {'noise': gn_chain, 'prior': gx_chain})
 
 
-@deprecate_kwarg({'iterations': 'num_iter'}, removed_version="23.02.00",
-                 deprecated_version="22.02.00")
 def richardson_lucy(image, psf, num_iter=50, clip=True, filter_epsilon=None):
     """Richardson-Lucy deconvolution.
 
diff --git a/python/cucim/src/cucim/skimage/restoration/j_invariant.py b/python/cucim/src/cucim/skimage/restoration/j_invariant.py
index e8a58dab2..1e19a92ca 100644
--- a/python/cucim/src/cucim/skimage/restoration/j_invariant.py
+++ b/python/cucim/src/cucim/skimage/restoration/j_invariant.py
@@ -89,14 +89,16 @@ def _generate_grid_slice(shape, *, offset, stride=3):
     return mask
 
 
-def _invariant_denoise(image, denoise_function, *, stride=4,
-                       masks=None, denoiser_kwargs=None):
+def denoise_invariant(image, denoise_function, *, stride=4, masks=None,
+                      denoiser_kwargs=None):
     """Apply a J-invariant version of `denoise_function`.
 
     Parameters
     ----------
-    image : ndarray
-        Input data to be denoised (converted using `img_as_float`).
+    image : ndarray ([M[, N[, ...P]][, C]) of ints, uints or floats
+        Input data to be denoised. `image` can be of any numeric type,
+        but it is cast into a ndarray of floats (using `img_as_float`) for the
+        computation of the denoised image.
     denoise_function : function
         Original denoising function.
     stride : int, optional
@@ -112,6 +114,32 @@ def _invariant_denoise(image, denoise_function, *, stride=4,
     -------
     output : ndarray
         Denoised image, of same shape as `image`.
+
+    Notes
+    -----
+    A denoising function is J-invariant if the prediction it makes for each
+    pixel does not depend on the value of that pixel in the original image.
+    The prediction for each pixel may instead use all the relevant information
+    contained in the rest of the image, which is typically quite significant.
+    Any function can be converted into a J-invariant one using a simple masking
+    procedure, as described in [1].
+
+    The pixel-wise error of a J-invariant denoiser is uncorrelated to the noise,
+    so long as the noise in each pixel is independent. Consequently, the average
+    difference between the denoised image and the oisy image, the
+    *self-supervised loss*, is the same as the difference between the denoised
+    image and the original clean image, the *ground-truth loss* (up to a
+    constant).
+
+    This means that the best J-invariant denoiser for a given image can be found
+    using the noisy data alone, by selecting the denoiser minimizing the self-
+    supervised loss.
+
+    References
+    ----------
+    .. [1] J. Batson & L. Royer. Noise2Self: Blind Denoising by
+       Self-Supervision, International Conference on Machine Learning,
+       p. 524-533 (2019).
     """
     image = img_as_float(image)
 
@@ -252,7 +280,7 @@ def calibrate_denoiser(image, denoise_function, denoise_parameters, *,
     best_parameters = parameters_tested[idx]
 
     best_denoise_function = functools.partial(
-        _invariant_denoise,
+        denoise_invariant,
         denoise_function=denoise_function,
         stride=stride,
         denoiser_kwargs=best_parameters,
@@ -303,7 +331,7 @@ def _calibrate_denoiser_search(image, denoise_function, denoise_parameters, *,
             multichannel = \
                 denoiser_kwargs.get('channel_axis', None) is not None
         if not approximate_loss:
-            denoised = _invariant_denoise(
+            denoised = denoise_invariant(
                 image, denoise_function,
                 stride=stride,
                 denoiser_kwargs=denoiser_kwargs
@@ -315,7 +343,7 @@ def _calibrate_denoiser_search(image, denoise_function, denoise_parameters, *,
             mask = _generate_grid_slice(image.shape[:spatialdims],
                                         offset=n_masks // 2, stride=stride)
 
-            masked_denoised = _invariant_denoise(
+            masked_denoised = denoise_invariant(
                 image, denoise_function,
                 masks=[mask],
                 denoiser_kwargs=denoiser_kwargs
diff --git a/python/cucim/src/cucim/skimage/restoration/tests/test_j_invariant.py b/python/cucim/src/cucim/skimage/restoration/tests/test_j_invariant.py
index 88d814cbb..0386d8950 100644
--- a/python/cucim/src/cucim/skimage/restoration/tests/test_j_invariant.py
+++ b/python/cucim/src/cucim/skimage/restoration/tests/test_j_invariant.py
@@ -10,7 +10,7 @@
 from cucim.skimage.data import binary_blobs
 from cucim.skimage.metrics import mean_squared_error as mse
 from cucim.skimage.restoration import calibrate_denoiser, denoise_tv_chambolle
-from cucim.skimage.restoration.j_invariant import _invariant_denoise
+from cucim.skimage.restoration.j_invariant import denoise_invariant
 from cucim.skimage.util import img_as_float, random_noise
 
 test_img = img_as_float(cp.asarray(camera()))
@@ -36,9 +36,9 @@ def _denoise_wavelet(image, rescale_sigma=True, **kwargs):
     )
 
 
-def test_invariant_denoise():
-    # denoised_img = _invariant_denoise(noisy_img, _denoise_wavelet)
-    denoised_img = _invariant_denoise(noisy_img, denoise_tv_chambolle)
+def test_denoise_invariant():
+    # denoised_img = denoise_invariant(noisy_img, _denoise_wavelet)
+    denoised_img = denoise_invariant(noisy_img, denoise_tv_chambolle)
 
     denoised_mse = mse(denoised_img, test_img)
     original_mse = mse(noisy_img, test_img)
@@ -46,8 +46,8 @@ def test_invariant_denoise():
 
 
 @pytest.mark.parametrize('dtype', [cp.float16, cp.float32, cp.float64])
-def test_invariant_denoise_color(dtype):
-    denoised_img_color = _invariant_denoise(
+def test_denoise_invariant_color(dtype):
+    denoised_img_color = denoise_invariant(
         noisy_img_color.astype(dtype),
         _denoise_wavelet,
         denoiser_kwargs=dict(channel_axis=-1),
@@ -59,8 +59,8 @@ def test_invariant_denoise_color(dtype):
     assert denoised_img_color.dtype == _supported_float_type(dtype)
 
 
-def test_invariant_denoise_3d():
-    denoised_img_3d = _invariant_denoise(noisy_img_3d, _denoise_wavelet)
+def test_denoise_invariant_3d():
+    denoised_img_3d = denoise_invariant(noisy_img_3d, _denoise_wavelet)
 
     denoised_mse = mse(denoised_img_3d, test_img_3d)
     original_mse = mse(noisy_img_3d, test_img_3d)
@@ -76,8 +76,8 @@ def test_calibrate_denoiser_extra_output():
         extra_output=True
     )
 
-    all_denoised = [_invariant_denoise(noisy_img, _denoise_wavelet,
-                                       denoiser_kwargs=denoiser_kwargs)
+    all_denoised = [denoise_invariant(noisy_img, _denoise_wavelet,
+                                      denoiser_kwargs=denoiser_kwargs)
                     for denoiser_kwargs in parameters_tested]
 
     ground_truth_losses = [float(mse(img, test_img)) for img in all_denoised]
diff --git a/python/cucim/src/cucim/skimage/restoration/tests/test_restoration.py b/python/cucim/src/cucim/skimage/restoration/tests/test_restoration.py
index 5d5ccd5a9..839f19f96 100644
--- a/python/cucim/src/cucim/skimage/restoration/tests/test_restoration.py
+++ b/python/cucim/src/cucim/skimage/restoration/tests/test_restoration.py
@@ -82,8 +82,7 @@ def test_unsupervised_wiener(dtype):
 
     psf = cp.asarray(psf, dtype=dtype)
     data = cp.asarray(data, dtype=dtype)
-    deconvolved, _ = restoration.unsupervised_wiener(data, psf,
-                                                     random_state=seed)
+    deconvolved, _ = restoration.unsupervised_wiener(data, psf, seed=seed)
     float_type = _supported_float_type(dtype)
     assert deconvolved.dtype == float_type
 
@@ -109,7 +108,7 @@ def test_unsupervised_wiener(dtype):
         user_params={"callback": lambda x: None,
                      "max_num_iter": 200,
                      "min_num_iter": 30},
-        random_state=seed,
+        seed=seed,
     )[0]
     assert deconvolved2.real.dtype == float_type
 
@@ -129,7 +128,8 @@ def test_unsupervised_wiener_deprecated_user_param():
     otf = uft.ir2tf(psf, data.shape, is_real=False)
     _, laplacian = uft.laplacian(2, data.shape)
     with expected_warnings(["`max_iter` is a deprecated key",
-                            "`min_iter` is a deprecated key"]):
+                            "`min_iter` is a deprecated key",
+                            "`random_state` is a deprecated argument name"]):
         restoration.unsupervised_wiener(
             data, otf, reg=laplacian, is_real=False,
             user_params={"max_iter": 200, "min_iter": 30}, random_state=5
@@ -175,17 +175,6 @@ def test_richardson_lucy():
     cp.testing.assert_allclose(deconvolved, np.load(path), rtol=1e-4)
 
 
-def test_richardson_lucy_deprecated_iterations_kwarg():
-    psf = np.ones((5, 5)) / 25
-    data = signal.convolve2d(cp.asnumpy(test_img), psf, 'same')
-    np.random.seed(0)
-    data += 0.1 * data.std() * np.random.standard_normal(data.shape)
-    data = cp.array(data)
-    psf = cp.array(psf)
-    with expected_warnings(["`iterations` is a deprecated argument"]):
-        restoration.richardson_lucy(data, psf, iterations=5)
-
-
 @pytest.mark.parametrize('dtype_image', [cp.float16, cp.float32, cp.float64])
 @pytest.mark.parametrize('dtype_psf', [cp.float32, cp.float64])
 @testing.with_requires("scikit-image>=0.18")
diff --git a/python/cucim/src/cucim/skimage/segmentation/_chan_vese.py b/python/cucim/src/cucim/skimage/segmentation/_chan_vese.py
index bab2f1887..84a7feb2b 100644
--- a/python/cucim/src/cucim/skimage/segmentation/_chan_vese.py
+++ b/python/cucim/src/cucim/skimage/segmentation/_chan_vese.py
@@ -8,37 +8,6 @@
 from .._vendored import pad
 
 
-@cp.fuse()
-def _fused_curvature(phi, x_start, x_end, y_start, y_end, ul, ur, ll, lr):
-    fy = (y_end - y_start) / 2.0
-    fx = (x_end - x_start) / 2.0
-    fyy = y_end + y_start - 2 * phi
-    fxx = x_end + x_start - 2 * phi
-    fxy = .25 * (lr + ul - ur - ll)
-    grad2 = fx**2 + fy**2
-    K = (fxx * fy**2 - 2 * fxy * fx * fy + fyy * fx**2)
-    K /= (grad2 * cp.sqrt(grad2) + 1e-8)
-    return K
-
-
-def _cv_curvature(phi):
-    """Returns the 'curvature' of a level set 'phi'.
-    """
-    P = pad(phi, 1, mode='edge')
-    y_start = P[:-2, 1:-1]
-    y_end = P[2:, 1:-1]
-    x_start = P[1:-1, :-2]
-    x_end = P[1:-1, 2:]
-
-    lower_right = P[2:, 2:]
-    lower_left = P[2:, :-2]
-    upper_right = P[:-2, 2:]
-    upper_left = P[:-2, :-2]
-    K = _fused_curvature(phi, x_start, x_end, y_start, y_end, upper_left,
-                         upper_right, lower_left, lower_right)
-    return K
-
-
 @cp.fuse()
 def _fused_variance_kernel1(eta, x_start, x_mid, x_end, y_start, y_mid, y_end):
     phixp = x_end - x_mid
@@ -102,6 +71,21 @@ def _fused_variance_kernel2(
 
 def _cv_calculate_variation(image, phi, mu, lambda1, lambda2, dt):
     """Returns the variation of level set 'phi' based on algorithm parameters.
+
+    This corresponds to equation (22) of the paper by Pascal Getreuer,
+    which computes the next iteration of the level set based on a current
+    level set.
+
+    A full explanation regarding all the terms is beyond the scope of the
+    present description, but there is one difference of particular import.
+    In the original algorithm, convergence is accelerated, and required
+    memory is reduced, by using a single array. This array, therefore, is a
+    combination of non-updated and updated values. If this were to be
+    implemented in python, this would require a double loop, where the
+    benefits of having fewer iterations would be outweided by massively
+    increasing the time required to perform each individual iteration. A
+    similar approach is used by Rami Cohen, and it is from there that the
+    C1-4 notation is taken.
     """
     eta = 1e-16
     P = pad(phi, 1, mode='edge')
@@ -183,17 +167,44 @@ def _cv_difference_from_average_term(image, Hphi, lambda_pos, lambda_neg):
     return out
 
 
+@cp.fuse()
+def _fused_edge_length(mu, phi, x_start, x_end, y_start, y_end):
+    fy = (y_end - y_start) / 2.0
+    fx = (x_end - x_start) / 2.0
+    grad2 = fx**2 + fy**2
+
+    del_phi = 1.0 / (1.0 + phi * phi)  # _cv_delta(phi)
+    K = mu * del_phi * cp.sqrt(grad2)
+    return K
+
+
 def _cv_edge_length_term(phi, mu):
-    """Returns the 'energy' contribution due to the length of the
-    edge between regions at each point, multiplied by a factor 'mu'.
+    """Returns the 'curvature' of a level set 'phi'.
     """
-    e = _cv_curvature(phi)
-    e *= mu
-    return e
+    P = pad(phi, 1, mode='edge')
+    y_start = P[:-2, 1:-1]
+    y_end = P[2:, 1:-1]
+    x_start = P[1:-1, :-2]
+    x_end = P[1:-1, 2:]
+
+    K = _fused_edge_length(mu, phi, x_start, x_end, y_start, y_end)
+    return K
 
 
 def _cv_energy(image, phi, mu, lambda1, lambda2):
     """Returns the total 'energy' of the current level set function.
+
+    This corresponds to equation (7) of the paper by Pascal Getreuer,
+    which is the weighted sum of the following:
+    (A) the length of the contour produced by the zero values of the
+    level set,
+    (B) the area of the "foreground" (area of the image where the
+    level set is positive),
+    (C) the variance of the image inside the foreground,
+    (D) the variance of the image outside of the foreground
+
+    Each value is computed for each pixel, and then summed. The weight
+    of (B) is set to 0 in this implementation.
     """
     H = _cv_heavyside(phi)
     avgenergy = _cv_difference_from_average_term(image, H, lambda1, lambda2)
diff --git a/python/cucim/src/cucim/skimage/segmentation/_join.py b/python/cucim/src/cucim/skimage/segmentation/_join.py
index 3744de969..6227a2fea 100644
--- a/python/cucim/src/cucim/skimage/segmentation/_join.py
+++ b/python/cucim/src/cucim/skimage/segmentation/_join.py
@@ -3,7 +3,7 @@
 from ..util._map_array import ArrayMap, map_array
 
 
-def join_segmentations(s1, s2):
+def join_segmentations(s1, s2, return_mapping: bool = False):
     """Return the join of the two input segmentations.
 
     The join J of S1 and S2 is defined as the segmentation in which two
@@ -14,11 +14,18 @@ def join_segmentations(s1, s2):
     ----------
     s1, s2 : numpy arrays
         s1 and s2 are label fields of the same shape.
+    return_mapping : bool, optional
+        If true, return mappings for joined segmentation labels to the original
+        labels.
 
     Returns
     -------
     j : numpy array
         The join segmentation of s1 and s2.
+    map_j_to_s1 : ArrayMap, optional
+        Mapping from labels of the joined segmentation j to labels of s1.
+    map_j_to_s2 : ArrayMap, optional
+        Mapping from labels of the joined segmentation j to labels of s2.
 
     Examples
     --------
@@ -38,11 +45,21 @@ def join_segmentations(s1, s2):
     if s1.shape != s2.shape:
         raise ValueError("Cannot join segmentations of different shape. "
                          f"s1.shape: {s1.shape}, s2.shape: {s2.shape}")
-    s1 = relabel_sequential(s1)[0]
-    s2 = relabel_sequential(s2)[0]
-    j = (s2.max() + 1) * s1 + s2
-    j = relabel_sequential(j)[0]
-    return j
+    s1_relabeled, _, backward_map1 = relabel_sequential(s1)
+    s2_relabeled, _, backward_map2 = relabel_sequential(s2)
+    factor = s2.max() + 1
+    j_initial = factor * s1_relabeled + s2_relabeled
+    j, _, map_j_to_j_initial = relabel_sequential(j_initial)
+    if not return_mapping:
+        return j
+    # Determine label mapping
+    labels_j = cp.unique(j_initial)
+    labels_s1_relabeled, labels_s2_relabeled = cp.divmod(labels_j, factor)
+    map_j_to_s1 = ArrayMap(map_j_to_j_initial.in_values,
+                           backward_map1[labels_s1_relabeled])
+    map_j_to_s2 = ArrayMap(map_j_to_j_initial.in_values,
+                           backward_map2[labels_s2_relabeled])
+    return j, map_j_to_s1, map_j_to_s2
 
 
 def relabel_sequential(label_field, offset=1):
diff --git a/python/cucim/src/cucim/skimage/segmentation/tests/test_join.py b/python/cucim/src/cucim/skimage/segmentation/tests/test_join.py
index af528073c..e68df7b8e 100644
--- a/python/cucim/src/cucim/skimage/segmentation/tests/test_join.py
+++ b/python/cucim/src/cucim/skimage/segmentation/tests/test_join.py
@@ -23,6 +23,12 @@ def test_join_segmentations():
                       [0, 5, 3, 2],
                       [4, 5, 5, 3]])
     assert_array_equal(j, j_ref)
+
+    # test correct mapping
+    j, m1, m2 = join_segmentations(s1, s2, return_mapping=True)
+    assert_array_equal(m1[j], s1)
+    assert_array_equal(m2[j], s2)
+
     # fmt: on
     # test correct exception when arrays are different shapes
     s3 = cp.array([[0, 0, 1, 1], [0, 2, 2, 1]])
diff --git a/python/cucim/src/cucim/skimage/transform/_geometric.py b/python/cucim/src/cucim/skimage/transform/_geometric.py
index bb6bde4bf..4da5b11f4 100644
--- a/python/cucim/src/cucim/skimage/transform/_geometric.py
+++ b/python/cucim/src/cucim/skimage/transform/_geometric.py
@@ -487,6 +487,39 @@ class EssentialMatrixTransform(FundamentalMatrixTransform):
     params : (3, 3) ndarray
         Essential matrix.
 
+    Examples
+    --------
+    >>> import cupy as cp
+    >>> from cucim.skimage import transform
+    >>>
+    >>> tform_matrix = transform.EssentialMatrixTransform(
+    ...     rotation=cp.eye(3), translation=cp.array([0, 0, 1])
+    ... )
+    >>> tform_matrix.params
+    array([[ 0., -1.,  0.],
+           [ 1.,  0.,  0.],
+           [ 0.,  0.,  0.]])
+    >>> src = cp.array([[ 1.839035, 1.924743],
+    ...                 [ 0.543582, 0.375221],
+    ...                 [ 0.47324 , 0.142522],
+    ...                 [ 0.96491 , 0.598376],
+    ...                 [ 0.102388, 0.140092],
+    ...                 [15.994343, 9.622164],
+    ...                 [ 0.285901, 0.430055],
+    ...                 [ 0.09115 , 0.254594]])
+    >>> dst = cp.array([[1.002114, 1.129644],
+    ...                 [1.521742, 1.846002],
+    ...                 [1.084332, 0.275134],
+    ...                 [0.293328, 0.588992],
+    ...                 [0.839509, 0.08729 ],
+    ...                 [1.779735, 1.116857],
+    ...                 [0.878616, 0.602447],
+    ...                 [0.642616, 1.028681]])
+    >>> tform_matrix.estimate(src, dst)
+    True
+    >>> tform_matrix.residuals(src, dst)
+    array([0.42455187, 0.01460448, 0.13847034, 0.12140951, 0.27759346,
+           0.32453118, 0.00210776, 0.26512283])
     """
 
     # CuPy Backend: if matrix is None cannot infer array module from it
@@ -874,14 +907,26 @@ class AffineTransform(ProjectiveTransform):
 
     Has the following form::
 
-        X = a0*x + a1*y + a2 =
-          = sx*x*cos(rotation) - sy*y*sin(rotation + shear) + a2
+        X = a0 * x + a1 * y + a2
+          =   sx * x * [cos(rotation) + tan(shear_y) * sin(rotation)]
+            - sy * y * [tan(shear_x) * cos(rotation) + sin(rotation)]
+            + translation_x
+
+        Y = b0 * x + b1 * y + b2
+          =   sx * x * [sin(rotation) - tan(shear_y) * cos(rotation)]
+            - sy * y * [tan(shear_x) * sin(rotation) - cos(rotation)]
+            + translation_y
+
+    where ``sx`` and ``sy`` are scale factors in the x and y directions.
 
-        Y = b0*x + b1*y + b2 =
-          = sx*x*sin(rotation) + sy*y*cos(rotation + shear) + b2
+    This is equivalent to applying the operations in the following order:
 
-    where ``sx`` and ``sy`` are scale factors in the x and y directions,
-    and the homogeneous transformation matrix is::
+    1. Scale
+    2. Shear
+    3. Rotate
+    4. Translate
+
+    The homogeneous transformation matrix is::
 
         [[a0  a1  a2]
          [b0  b1  b2]
@@ -909,11 +954,12 @@ class AffineTransform(ProjectiveTransform):
         .. versionadded:: 0.17
            Added support for supplying a single scalar value.
     rotation : float, optional
-        Rotation angle in counter-clockwise direction as radians. Only
-        available for 2D.
-    shear : float, optional
-        Shear angle in counter-clockwise direction as radians. Only available
-        for 2D.
+        Rotation angle, clockwise, as radians. Only available for 2D.
+    shear : float or 2-tuple of float, optional
+        The x and y shear angles, clockwise, by which these axes are
+        rotated around the origin [2].
+        If a single value is given, take that to be the x shear angle, with
+        the y angle remaining 0. Only available in 2D.
     translation : (tx, ty) as ndarray, list or tuple, optional
         Translation parameters. Only available for 2D.
     dimensionality : int, optional
@@ -929,6 +975,39 @@ class AffineTransform(ProjectiveTransform):
     ------
     ValueError
         If both ``matrix`` and any of the other parameters are provided.
+
+    Examples
+    --------
+    >>> import cupy as cp
+    >>> from cucim.skimage import transform
+    >>> from skimage import data
+    >>> img = cp.array(data.astronaut())
+
+    Define source and destination points:
+
+    >>> src = cp.array([[150, 150],
+    ...                 [250, 100],
+    ...                 [150, 200]])
+    >>> dst = cp.array([[200, 200],
+    ...                 [300, 150],
+    ...                 [150, 400]])
+
+    Estimate the transformation matrix:
+
+    >>> tform = transform.AffineTransform()
+    >>> tform.estimate(src, dst)
+    True
+
+    Apply the transformation:
+
+    >>> warped = transform.warp(img, inverse_map=tform.inverse)
+
+    References
+    ----------
+    .. [1] Wikipedia, "Affine transformation",
+           https://en.wikipedia.org/wiki/Affine_transformation#Image_transformation
+    .. [2] Wikipedia, "Shear mapping",
+           https://en.wikipedia.org/wiki/Shear_mapping
     """
 
     def __init__(self, matrix=None, scale=None, rotation=None, shear=None,
@@ -968,17 +1047,27 @@ def __init__(self, matrix=None, scale=None, rotation=None, shear=None,
             else:
                 sx, sy = scale
 
-            # fmt: off
-            self.params = np.array(
-                [
-                    [sx * _cos(rotation), -sy * _sin(rotation + shear), 0],  # NOQA
-                    [sx * _sin(rotation),  sy * _cos(rotation + shear), 0],  # NOQA
-                    [                  0,                            0, 1],  # NOQA
-                ]
+            if np.isscalar(shear):
+                shear_x, shear_y = (shear, 0)
+            else:
+                shear_x, shear_y = shear
+
+            a0 = sx * (
+                math.cos(rotation) + math.tan(shear_y) * math.sin(rotation)
             )
-            # fmt: on
-            self.params[0:2, 2] = translation
-            self.params = xp.asarray(self.params)
+            a1 = -sy * (
+                math.tan(shear_x) * math.cos(rotation) + math.sin(rotation)
+            )
+            a2 = translation[0]
+
+            b0 = sx * (
+                math.sin(rotation) - math.tan(shear_y) * math.cos(rotation)
+            )
+            b1 = -sy * (
+                math.tan(shear_x) * math.sin(rotation) - math.cos(rotation)
+            )
+            b2 = translation[1]
+            self.params = xp.array([[a0, a1, a2], [b0, b1, b2], [0, 0, 1]])
         else:
             # default to an identity transform
             self.params = xp.eye(dimensionality + 1)
@@ -986,9 +1075,14 @@ def __init__(self, matrix=None, scale=None, rotation=None, shear=None,
     @property
     def scale(self):
         xp = cp.get_array_module(self.params)
-        return xp.sqrt(xp.sum(self.params * self.params, axis=0))[
-            : self.dimensionality
-        ]
+        if self.dimensionality != 2:
+            return xp.sqrt(xp.sum(self.params * self.params, axis=0))[
+                : self.dimensionality
+            ]
+        else:
+            ss = xp.sum(self.params * self.params, axis=0)
+            ss[1] = ss[1] / (math.tan(self.shear) ** 2 + 1)
+            return xp.sqrt(ss)[:self.dimensionality]
 
     @property
     def rotation(self):
@@ -1066,7 +1160,13 @@ def estimate(self, src, dst):
 
         # find affine mapping from source positions to destination
         self.affines = []
-        for tri in xp.asarray(self._tesselation.vertices):
+
+        try:
+            tesselation_simplices = self._tesselation.simplices
+        except AttributeError:
+            # vertices is deprecated and scheduled for removal in SciPy 1.11
+            tesselation_simplices = self._tesselation.vertices
+        for tri in xp.asarray(tesselation_simplices):
             affine = AffineTransform(dimensionality=ndim)
             ok &= affine.estimate(src[tri, :], dst[tri, :])
             self.affines.append(affine)
@@ -1077,7 +1177,12 @@ def estimate(self, src, dst):
         self._inverse_tesselation = spatial.Delaunay(cp.asnumpy(dst))
         # find affine mapping from source positions to destination
         self.inverse_affines = []
-        for tri in xp.asarray(self._inverse_tesselation.vertices):
+        try:
+            inv_tesselation_simplices = self._inverse_tesselation.simplices
+        except AttributeError:
+            # vertices is deprecated and scheduled for removal in SciPy 1.11
+            inv_tesselation_simplices = self._inverse_tesselation.vertices
+        for tri in xp.asarray(inv_tesselation_simplices):
             affine = AffineTransform(dimensionality=ndim)
             ok &= affine.estimate(dst[tri, :], src[tri, :])
             self.inverse_affines.append(affine)
@@ -1112,7 +1217,12 @@ def __call__(self, coords):
         # coordinates outside of mesh
         out[simplex == -1, :] = -1
 
-        for index in range(len(self._tesselation.vertices)):
+        try:
+            tesselation_simplices = self._tesselation.simplices
+        except AttributeError:
+            # vertices is deprecated and scheduled for removal in SciPy 1.11
+            tesselation_simplices = self._tesselation.vertices
+        for index in range(len(tesselation_simplices)):
             # affine transform for triangle
             affine = self.affines[index]
             # all coordinates within triangle
@@ -1150,7 +1260,12 @@ def inverse(self, coords):
         # coordinates outside of mesh
         out[simplex == -1, :] = -1
 
-        for index in range(len(self._inverse_tesselation.vertices)):
+        try:
+            inv_tesselation_simplices = self._inverse_tesselation.simplices
+        except AttributeError:
+            # vertices is deprecated and scheduled for removal in SciPy 1.11
+            inv_tesselation_simplices = self._inverse_tesselation.vertices
+        for index in range(len(inv_tesselation_simplices)):
             # affine transform for triangle
             affine = self.inverse_affines[index]
             # all coordinates within triangle
@@ -1244,7 +1359,7 @@ class EuclideanTransform(ProjectiveTransform):
     matrix : (D+1, D+1) ndarray, optional
         Homogeneous transformation matrix.
     rotation : float or sequence of float, optional
-        Rotation angle in counter-clockwise direction as radians. If given as
+        Rotation angle, clockwise, as radians. If given as
         a vector, it is interpreted as Euler rotation angles [1]_. Only 2D
         (single rotation) and 3D (Euler rotations) values are supported. For
         higher dimensions, you must provide or estimate the transformation
@@ -1390,7 +1505,7 @@ class SimilarityTransform(EuclideanTransform):
     scale : float, optional
         Scale factor. Implemented only for 2D and 3D.
     rotation : float, optional
-        Rotation angle in counter-clockwise direction as radians.
+        Rotation angle, clockwise, as radians.
         Implemented only for 2D and 3D. For 3D, this is given in XZX Euler
         angles.
     translation : (dim,) ndarray-like, optional
diff --git a/python/cucim/src/cucim/skimage/transform/_warps.py b/python/cucim/src/cucim/skimage/transform/_warps.py
index 88a395591..343f1f0f7 100644
--- a/python/cucim/src/cucim/skimage/transform/_warps.py
+++ b/python/cucim/src/cucim/skimage/transform/_warps.py
@@ -166,9 +166,6 @@ def resize(image, output_shape, order=None, mode='reflect', cval=0, clip=None,
     if clip is None:
         clip = True if order > 1 else False
 
-    # Save input value range for clip
-    img_bounds = (image.min(), image.max()) if clip else None
-
     # Translate modes used by np.pad to those used by scipy.ndimage
     ndi_mode = _to_ndimage_mode(mode)
     if anti_aliasing:
@@ -192,13 +189,15 @@ def resize(image, output_shape, order=None, mode='reflect', cval=0, clip=None,
         _ndi_mode = {'grid-constant': 'constant', 'grid-wrap':'wrap'}.get(ndi_mode, ndi_mode)  # noqa
         # keep ndi.gaussian_filter rather than cucim.skimage.filters.gaussian
         # to avoid undesired dtype coercion
-        image = ndi.gaussian_filter(image, anti_aliasing_sigma, cval=cval,
-                                    mode=_ndi_mode)
+        filtered = ndi.gaussian_filter(image, anti_aliasing_sigma, cval=cval,
+                                       mode=_ndi_mode)
+    else:
+        filtered = image
 
     zoom_factors = [1 / f for f in factors]
-    out = ndi.zoom(image, zoom_factors, order=order, mode=ndi_mode,
+    out = ndi.zoom(filtered, zoom_factors, order=order, mode=ndi_mode,
                    cval=cval, grid_mode=True)
-    _clip_warp_output(img_bounds, out, mode, cval, order, clip)
+    _clip_warp_output(image, out, mode, cval, order, clip)
     return out
 
 
@@ -821,7 +820,7 @@ def _clip_warp_output(input_image, output_image, mode, cval, order, clip):
     cp.clip(output_image, min_val, max_val, out=output_image)
 
 
-def warp(image, inverse_map, map_args={}, output_shape=None, order=None,
+def warp(image, inverse_map, map_args=None, output_shape=None, order=None,
          mode='constant', cval=0., clip=None, preserve_range=False):
     """Warp an image according to a given coordinate transformation.
 
@@ -963,6 +962,9 @@ def warp(image, inverse_map, map_args={}, output_shape=None, order=None,
     >>> warped = warp(cube, coords)
 
     """  # noqa
+    if map_args is None:
+        map_args = {}
+
     if image.size == 0:
         raise ValueError("Cannot warp empty image with dimensions",
                          image.shape)
diff --git a/python/cucim/src/cucim/skimage/transform/tests/test_geometric.py b/python/cucim/src/cucim/skimage/transform/tests/test_geometric.py
index 2b2891542..d9aee75db 100644
--- a/python/cucim/src/cucim/skimage/transform/tests/test_geometric.py
+++ b/python/cucim/src/cucim/skimage/transform/tests/test_geometric.py
@@ -269,6 +269,38 @@ def test_affine_init():
                               AffineTransform(scale=(0.5, 0.5)).scale)
 
 
+@pytest.mark.parametrize("xp", [np, cp])
+def test_affine_shear(xp):
+    shear = 0.1
+    # expected horizontal shear transform
+    cx = -np.tan(shear)
+    # fmt: off
+    expected = xp.array([
+        [1, cx, 0],
+        [0,  1, 0],  # noqa: E241
+        [0,  0, 1],  # noqa: E241
+    ])
+    # fmt: on
+
+    tform = AffineTransform(shear=shear, xp=xp)
+    xp.testing.assert_array_almost_equal(tform.params, expected)
+
+    shear = (1.2, 0.8)
+    # expected x, y shear transform
+    cx = -np.tan(shear[0])
+    cy = -np.tan(shear[1])
+    # fmt: off
+    expected = xp.array([
+        [ 1, cx, 0],  # noqa: E201
+        [cy,  1, 0],  # noqa: E241
+        [ 0,  0, 1],  # noqa: E241, E201
+    ])
+    # fmt: on
+
+    tform = AffineTransform(shear=shear, xp=xp)
+    xp.testing.assert_array_almost_equal(tform.params, expected)
+
+
 def test_piecewise_affine():
     tform = PiecewiseAffineTransform()
     assert tform.estimate(SRC, DST)
diff --git a/python/cucim/src/cucim/skimage/transform/tests/test_warps.py b/python/cucim/src/cucim/skimage/transform/tests/test_warps.py
index de7bbb06c..dbb47738e 100644
--- a/python/cucim/src/cucim/skimage/transform/tests/test_warps.py
+++ b/python/cucim/src/cucim/skimage/transform/tests/test_warps.py
@@ -472,10 +472,12 @@ def test_resize_clip(order, preserve_range, anti_aliasing, dtype):
     x = cp.ones((5, 5), dtype=dtype)
     if dtype == cp.uint8:
         x *= 255
+    else:
+        x[0, 0] = cp.NaN
     resized = resize(x, (3, 3), order=order, preserve_range=preserve_range,
                      anti_aliasing=anti_aliasing)
 
-    assert abs(float(resized.max()) - expected_max) < 1e-14
+    assert abs(float(cp.nanmax(resized)) - expected_max) < 1e-14
 
 
 @pytest.mark.parametrize('dtype', [cp.float16, cp.float32, cp.float64])