Soft Color Morphology

Color extension of Mathematical Morphology using Fuzzy Operators to deal with grayscale values.

Bibiloni, P., González-Hidalgo, M., & Massanet, S. (2018). 
Soft Color Morphology: A Fuzzy Approach for Multivariate Images. 
Journal of Mathematical Imaging and Vision, 1-17.

Getting Started

Soft color morphology provide two operators, erosion and dilation, that can be combined to form other complex mathematical morphology operators. It admits non-binary structuring elements.

To install, type in terminal:

pip install softcolor

To use in natural images, we recommend using the CIELab color space:

import numpy as np
from skimage import data
from softcolor.morphology import MorphologyInCIELab

morphology = MorphologyInCIELab()
structuring_element = np.ones(shape=(5, 1), dtype='float32')

image = data.astronaut()
image_eroded = morphology.erosion(image, structuring_element)
image_dilated = morphology.dilation(image, structuring_element)

The soft color morphology operators depend on a fuzzy conjunction and a fuzzy implication function. They can be provided as:

from skimage import data
from skimage.morphology import disk
from softcolor.morphology import MorphologyInCIELab, soften_structuring_element
from softcolor.aggregation_functions import conjunction_min, implication_godel

morphology = MorphologyInCIELab(
    conjunction=conjunction_min,
    fuzzy_implication_function=implication_godel)
structuring_element = soften_structuring_element(disk(5))

image = data.astronaut()
image_eroded = morphology.opening(image, structuring_element)
image_dilated = morphology.closing(image, structuring_element)

Some operations (e.g. top-hat, inpainting) also depend on a measure of dissimilarity between images (such as the pixel-wise Euclidean distance) and a method to combine images (such as the standard average).

To apply the soft color operators with generic color spaces, use the class softcolor.morphology.BaseMorphology.

More examples

Browse the examples folder to check out other morphological operations.

License

This project is licensed under the MIT License - see the LICENSE file for details.