Source code for shapevolve.error_metrics

"""Defines error metrics that can be used to evaluate similarity between two images."""

import skimage.metrics  # For similarity measurement


[docs]def mean_squared_error(image, source): """Calculates the mean squared error between two images. :param image: The constructed image. :param source: The original image. :type image: ndarray :type source: ndarray :return: The mean squared error between the two images. :rtype: float """ return skimage.metrics.mean_squared_error(image, source)
[docs]def structural_similarity_error(image, source): """Calculates the opposite of the structural similarity between two images. :param image: The constructed image. :param source: The original image. :type image: ndarray :type source: ndarray :return: The opposite of the structural similarity between the two images. :rtype: float """ return -1 * skimage.metrics.structural_similarity(image, source, multichannel=True)