Image Processing¶
These are explicit differentiable counterparts to
scipy.ndimage.
Use the compatibility table for exact dynamic,
staged, and serialized coverage.
ndimage ¶
Traceable counterparts to frequently used scipy.ndimage operations.
gaussian_filter ¶
gaussian_filter(
input: object,
sigma: object,
order: object = 0,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
truncate: object = 4.0,
*,
radius: object = None,
axes: object = None,
) -> object
Apply a multidimensional Gaussian filter with exact boundary adjoints.
gaussian_filter1d ¶
gaussian_filter1d(
input: object,
sigma: object,
axis: object = -1,
order: object = 0,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
truncate: object = 4.0,
*,
radius: object = None,
) -> object
Apply a one-dimensional Gaussian filter along axis.
uniform_filter ¶
uniform_filter(
input: object,
size: object = 3,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
origin: object = 0,
*,
axes: object = None,
) -> object
Apply a multidimensional uniform filter.
uniform_filter1d ¶
uniform_filter1d(
input: object,
size: object,
axis: object = -1,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
origin: object = 0,
) -> object
Apply a one-dimensional uniform filter along axis.
convolve ¶
convolve(
input: object,
weights: object,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
origin: object = 0,
*,
axes: object = None,
) -> object
Multidimensional convolution with differentiable input and weights.
correlate ¶
correlate(
input: object,
weights: object,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
origin: object = 0,
*,
axes: object = None,
) -> object
Multidimensional correlation with differentiable input and weights.
convolve1d ¶
convolve1d(
input: object,
weights: object,
axis: object = -1,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
origin: object = 0,
) -> object
One-dimensional convolution with differentiable input and weights.
correlate1d ¶
correlate1d(
input: object,
weights: object,
axis: object = -1,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
origin: object = 0,
) -> object
One-dimensional correlation with differentiable input and weights.
laplace ¶
laplace(
input: object,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
*,
axes: object = None,
) -> object
Apply the discrete multidimensional Laplace operator.
gaussian_laplace ¶
gaussian_laplace(
input: object,
sigma: object,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
*,
axes: object = None,
**kwargs: object,
) -> object
Apply a Laplacian of Gaussian filter.
sobel ¶
sobel(
input: object,
axis: object = -1,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
) -> object
Calculate an axis-specific Sobel filter.
prewitt ¶
prewitt(
input: object,
axis: object = -1,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
) -> object
Calculate an axis-specific Prewitt filter.
maximum_filter ¶
maximum_filter(
input: object,
size: object = None,
footprint: object = None,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
origin: object = 0,
*,
axes: object = None,
) -> object
Calculate a multidimensional maximum filter with symmetric tie gradients.
minimum_filter ¶
minimum_filter(
input: object,
size: object = None,
footprint: object = None,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
origin: object = 0,
*,
axes: object = None,
) -> object
Calculate a multidimensional minimum filter with symmetric tie gradients.
maximum_filter1d ¶
maximum_filter1d(
input: object,
size: object,
axis: object = -1,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
origin: object = 0,
) -> object
Calculate a one-dimensional maximum filter.
minimum_filter1d ¶
minimum_filter1d(
input: object,
size: object,
axis: object = -1,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
origin: object = 0,
) -> object
Calculate a one-dimensional minimum filter.
grey_dilation ¶
grey_dilation(
input: object,
size: object = None,
footprint: object = None,
structure: object = None,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
origin: object = 0,
*,
axes: object = None,
) -> object
Calculate a greyscale dilation with symmetric tie gradients.
grey_erosion ¶
grey_erosion(
input: object,
size: object = None,
footprint: object = None,
structure: object = None,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
origin: object = 0,
*,
axes: object = None,
) -> object
Calculate a greyscale erosion with symmetric tie gradients.
median_filter ¶
median_filter(
input: object,
size: object = None,
footprint: object = None,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
origin: object = 0,
*,
axes: object = None,
) -> object
Calculate a multidimensional median filter.
rank_filter ¶
rank_filter(
input: object,
rank: object,
size: object = None,
footprint: object = None,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
origin: object = 0,
*,
axes: object = None,
) -> object
Calculate a multidimensional rank filter.
percentile_filter ¶
percentile_filter(
input: object,
percentile: object,
size: object = None,
footprint: object = None,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
origin: object = 0,
*,
axes: object = None,
) -> object
Calculate a multidimensional percentile filter.
grey_opening ¶
grey_opening(
input: object,
size: object = None,
footprint: object = None,
structure: object = None,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
origin: object = 0,
*,
axes: object = None,
) -> object
Apply greyscale erosion followed by greyscale dilation.
grey_closing ¶
grey_closing(
input: object,
size: object = None,
footprint: object = None,
structure: object = None,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
origin: object = 0,
*,
axes: object = None,
) -> object
Apply greyscale dilation followed by greyscale erosion.
morphological_gradient ¶
morphological_gradient(
input: object,
size: object = None,
footprint: object = None,
structure: object = None,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
origin: object = 0,
*,
axes: object = None,
) -> object
Calculate the difference between greyscale dilation and erosion.
morphological_laplace ¶
morphological_laplace(
input: object,
size: object = None,
footprint: object = None,
structure: object = None,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
origin: object = 0,
*,
axes: object = None,
) -> object
Calculate the morphological Laplace operator.
white_tophat ¶
white_tophat(
input: object,
size: object = None,
footprint: object = None,
structure: object = None,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
origin: object = 0,
*,
axes: object = None,
) -> object
Calculate the difference between the input and its greyscale opening.
black_tophat ¶
black_tophat(
input: object,
size: object = None,
footprint: object = None,
structure: object = None,
output: object = None,
mode: object = "reflect",
cval: object = 0.0,
origin: object = 0,
*,
axes: object = None,
) -> object
Calculate the difference between greyscale closing and the input.
$ █