colourlab.image module¶

image: Colour image, part of the colourlab package

This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along with this program. If not, see <http://www.gnu.org/licenses/>.

class colourlab.image.Image(sp, ndata)[source]¶

Bases: colourlab.data.Points

Subclass of data.Points specifically for image shaped data.

anisotropic_diffusion(sp, nit, dpsi_dlambda1=None, dpsi_dlambda2=None, dt=0.25, linear=True, g_func=None, constraint=None)[source]¶

Compute the anisotropic diffusion of the image.

Parameters:	sp (space.Space) – The colour space in which to perform the diffusion nit (int) – Number of iterations dpsi_dlambda1 (func) – The diffusion supression function for the first eigenvalue of the structure tensor. If None, use Perona and Malik’s inverse square with kappa = 1e-2. dpsi_dlambda2 (func) – Same for the second eigenvalue. If None use dpsi_dlambda1 linear (bool) – Linear anisotropic diffusion if true g_func (func) – Function computing the metric tensor to use. If not given, uses Euclidean in the given space constraint (func) – Function returning the constrained image (e.g., gamut or internal boundaries, e.g. CFAs or fixed domains for inpainting) in sp
Returns:	space.Image
Return type:	the resulting anisotropically diffused image

c2g_diffusion(sp, nit, g_func=None, l_minus=True, scale=1, dt=0.25, dpsi_dlambda1=None, dpsi_dlambda2=None)[source]¶

Convert colour image to greyscale using linear anisotropic diffusion

Parameters:	sp (Space) – Colour space in which to perform the numerical computations nit (int) – Number of iterations to compute g (Tensors) – The colour metric tensor. If not given, use Euclidean l_minus (bool) – Use lambda_minus in the computation of the gradient scale (float) – Distance from black to white according to metric dt (float) – Time step dpsi_dlambda1 (func) – The diffusion supression function for the first eigenvalue of the structure tensor. If None, use Perona and Malik’s inverse square with kappa = 1e-2. dpsi_dlambda2 (func) – Same for the second eigenvalue. If None use dpsi_dlambda1
Returns:	grey_image – Greyscale image (range 0–1)
Return type:	ndarray

diffusion_tensor(sp, dpsi_dlambda1=None, dpsi_dlambda2=None, g_func=None, diff=(array([[ 0. , -0.5, 0. ], [ 0. , 0. , 0. ], [ 0. , 0.5, 0. ]]), array([[ 0. , 0. , 0. ], [-0.5, 0. , 0.5], [ 0. , 0. , 0. ]])), grey=None)[source]¶

Compute the diffusion tensor coefficients for the image point set

Assumes (for now) that the underlying data constitutes an image, i.e., is on the shape M x N x 3.

Parameters ———-100100 sp : Space

The space in which to perform the computations

dpsi_dlambda1 : func: The diffusion supression function for the first eigenvalue of the structure tensor. If None, use Perona and Malik’s inverse square with kappa = 1e-2.
dpsi_dlambda2 : func: Same for the second eigenvalue. If None use dpsi_dlambda1
g_func : func: Function computing the metric tensor to use. If not given, uses Euclidean in the given space

Returns:	d11 (ndarray) – The d11 component of the structure tensor of the image data. d12 (ndarray) – The d12 component of the structure tensor of the image data. d22 (ndarray) – The d22 component of the structure tensor of the image data.

gradient(sp, diff=(array([[ 0. , -0.5, 0. ], [ 0. , 0. , 0. ], [ 0. , 0.5, 0. ]]), array([[ 0. , 0. , 0. ], [-0.5, 0. , 0.5], [ 0. , 0. , 0. ]])))[source]¶

Compute the gradient components in the given colour space

Parameters:

sp (colour.space.Space) – The colour space
diff (tuple) – Tuple with the two filters for the derivatives. Defaults to centered differences

Returns:

image.data.Vector (The i component of the gradient as a colour vector)
image.data.Vector (The j component of the gradient as a colour vector)

stress(sp_in, sp_out=None, ns=3, nit=5, R=0)[source]¶

Compute STRESS in the given colour space.

Parameters:	sp_in (space.Space) – The input colour space. sp_out (space.Space) – The colour space for the interpretation of the result. If None, it is taken to be the same as sp_in. ns (int) – Number of sample points. nit (int) – Number of iterations. R (int) – Radius in pixels. If R=0, the diagonal of the image is used.
Returns:	The resulting STRESS image.
Return type:	image.Image

structure_tensor(sp, g_func=None, diff=(array([[ 0. , -0.5, 0. ], [ 0. , 0. , 0. ], [ 0. , 0.5, 0. ]]), array([[ 0. , 0. , 0. ], [-0.5, 0. , 0.5], [ 0. , 0. , 0. ]])), grey=None)[source]¶

Return the structure tensor of the underlying data image point set

Assumes (for now) that the underlying data constitutes an image, i.e., is on the shape M x N x 3. Note that the returned eigenvectors are not oriented in a particular way (+- pi).

Parameters:

sp (Space) – The space in which to perform the computations
g_func (func) – Function computing the metric tensor to use. If not given, uses Euclidean in the given space
diff (tuple) – Tuple with the two filters for the derivatives. Defaults to centered differences
grey (ndarray Grey scale image for orientation of lightness) – gradient. If not present, use CIELAB L* channel

Returns:

s11 (ndarray) – The s11 component of the structure tensor of the image data.
s12 (ndarray) – The s12 component of the structure tensor of the image data.
s22 (ndarray) – The s22 component of the structure tensor of the image data.
lambda1 (ndarray) – The first eigenvalue of the structure tensor
lambda2 (ndarray) – The second eigenvalue of the structure tensor
e1i (ndarray) – The first component of the first eigenvector
e1j (ndarray) – The second component of the first eigenvector
e2i (ndarray) – The first component of the second eigenvector
e2j (ndarray) – The second component of the second eigenvector