Module `circle_evolution.fitness`

Fitness functions responsible for quantifying a specie's performance.

You can make your own fitness functions or use the avaible one's. For examples on making your own fitness functions, please check the available functions in the source code.

Expand source code

"""Fitness functions responsible for quantifying a specie's performance.

You can make your own fitness functions or use the avaible one's. For
examples on making your own fitness functions, please check the available
functions in the source code.
"""

import numpy as np

from skimage.metrics import structural_similarity as ss


class Fitness:
    """Base fitness class.

    The Fitness class is responsible for scoring species. You can
    make your own fitness classes by deriving from this base class.

    Attributes:
        target (np.ndarray): target image array.
    """

    def __init__(self, target):
        """Initializes Fitness Class.

        Args:
            target (np.ndarray): target image array.
        """
        self.target = target

    def score(self, phenotype):
        """Score a Specie.

        Args:
            phenotype (np.ndarray): specie image array.

        Returns:
            float value, bigger number means better performance.

        Raises:
            NotImplementedError: if it has not been overwritten.
        """
        raise NotImplementedError


class MSEFitness(Fitness):
    """Mean Squared Error Fitness

    See: https://en.wikipedia.org/wiki/Mean_squared_error.
    """
    def __init__(self, *args, **kwargs):
        super().__init__(*args, **kwargs)
        self._max_error = (np.square((1 - (self.target >= 127)) * 255 - self.target)).mean(axis=None)

    def score(self, phenotype):
        """Calculates Mean Square Error Fitness for a specie"""
        fit = (np.square(phenotype - self.target)).mean(axis=None)
        fit = (self._max_error - fit) / self._max_error
        return fit


class SSFitness(Fitness):
    """Structural Similarity Fitness

    See: https://en.wikipedia.org/wiki/Structural_similarity.
    """

    def score(self, phenotype):
        """Calculates SS Fitness for a specie"""
        fit = ss(phenotype, self.target)
        return fit

Classes

class Fitness (target)

Base fitness class.

The Fitness class is responsible for scoring species. You can make your own fitness classes by deriving from this base class.

Attributes

target : np.ndarray: target image array.

Initializes Fitness Class.

Args

target : np.ndarray: target image array.

Expand source code

class Fitness:
    """Base fitness class.

    The Fitness class is responsible for scoring species. You can
    make your own fitness classes by deriving from this base class.

    Attributes:
        target (np.ndarray): target image array.
    """

    def __init__(self, target):
        """Initializes Fitness Class.

        Args:
            target (np.ndarray): target image array.
        """
        self.target = target

    def score(self, phenotype):
        """Score a Specie.

        Args:
            phenotype (np.ndarray): specie image array.

        Returns:
            float value, bigger number means better performance.

        Raises:
            NotImplementedError: if it has not been overwritten.
        """
        raise NotImplementedError

Methods

def score(self, phenotype)

Score a Specie.

Args

phenotype : np.ndarray: specie image array.

Returns

float value, bigger number means better performance.

Raises

NotImplementedError: if it has not been overwritten.

Expand source code

def score(self, phenotype):
    """Score a Specie.

    Args:
        phenotype (np.ndarray): specie image array.

    Returns:
        float value, bigger number means better performance.

    Raises:
        NotImplementedError: if it has not been overwritten.
    """
    raise NotImplementedError

class MSEFitness (*args, **kwargs)

Mean Squared Error Fitness

See: https://en.wikipedia.org/wiki/Mean_squared_error.

Initializes Fitness Class.

Args

target : np.ndarray: target image array.

Expand source code

class MSEFitness(Fitness):
    """Mean Squared Error Fitness

    See: https://en.wikipedia.org/wiki/Mean_squared_error.
    """
    def __init__(self, *args, **kwargs):
        super().__init__(*args, **kwargs)
        self._max_error = (np.square((1 - (self.target >= 127)) * 255 - self.target)).mean(axis=None)

    def score(self, phenotype):
        """Calculates Mean Square Error Fitness for a specie"""
        fit = (np.square(phenotype - self.target)).mean(axis=None)
        fit = (self._max_error - fit) / self._max_error
        return fit

Ancestors

Fitness

Methods

def score(self, phenotype)

Calculates Mean Square Error Fitness for a specie

Expand source code

def score(self, phenotype):
    """Calculates Mean Square Error Fitness for a specie"""
    fit = (np.square(phenotype - self.target)).mean(axis=None)
    fit = (self._max_error - fit) / self._max_error
    return fit

class SSFitness (target)

Structural Similarity Fitness

See: https://en.wikipedia.org/wiki/Structural_similarity.

Initializes Fitness Class.

Args

target : np.ndarray: target image array.

Expand source code

class SSFitness(Fitness):
    """Structural Similarity Fitness

    See: https://en.wikipedia.org/wiki/Structural_similarity.
    """

    def score(self, phenotype):
        """Calculates SS Fitness for a specie"""
        fit = ss(phenotype, self.target)
        return fit

Ancestors

Fitness

Methods

def score(self, phenotype)

Calculates SS Fitness for a specie

Expand source code

def score(self, phenotype):
    """Calculates SS Fitness for a specie"""
    fit = ss(phenotype, self.target)
    return fit