From 802cd208dbea7743a779b058478b6bb061e0d62d Mon Sep 17 00:00:00 2001
From: Rafael Zimmer <rzimmerdev@gmail.com>
Date: Tue, 18 Oct 2022 22:01:06 -0300
Subject: [PATCH 01/17] Create haralick_descriptors

---
 computer_vision/haralick_descriptors | 245 +++++++++++++++++++++++++++
 1 file changed, 245 insertions(+)
 create mode 100644 computer_vision/haralick_descriptors

diff --git a/computer_vision/haralick_descriptors b/computer_vision/haralick_descriptors
new file mode 100644
index 000000000000..6f8d6bd66b80
--- /dev/null
+++ b/computer_vision/haralick_descriptors
@@ -0,0 +1,245 @@
+def rmse(original: np.ndarray, reference: np.ndarray) -> float:
+    """Simple implementation of Root Mean Squared Error for two N dimensional numpy arrays."""
+    return np.sqrt(((original - reference) ** 2).mean())
+
+
+def normalize_image(image: np.ndarray, cap: float = 255) -> np.ndarray:
+    """Normalizes image in Numpy 2D array format, between ranges 0-cap, as to fit uint8 type.
+
+    Args:
+        image: 2D numpy array representing image as matrix, with values in any range
+        cap: Maximum cap amount for normalization
+    Returns:
+        return 2D numpy array of type uint8, corresponding to limited range matrix
+    """
+    normalized = (image - np.min(image)) / (np.max(image) - np.min(image)) * cap
+    return normalized.astype(np.uint8)
+
+
+def normalize_array(array, cap: float = 1):
+    """Normalizes a 1D array, between ranges 0-cap.
+
+    Args:
+        array: List containing values to be normalized between cap range.
+        cap: Maximum cap amount for normalization.
+    Returns:
+        return 1D numpy array , corresponding to limited range array
+    """
+    normalized = (array - np.min(array)) / (np.max(array) - np.min(array)) * cap
+    return normalized
+
+
+def euclidean(point_1: np.ndarray, point_2: np.ndarray):
+    """Simple method for calculating the euclidean distance between two points, with type np.ndarray."""
+    return np.sqrt(np.sum(np.square(point_1 - point_2)))
+
+
+def grayscale(image: np.ndarray) -> np.ndarray:
+    """Uses luminance weights to transform RGB channel to greyscale, by
+    taking the dot product between the channel and the weights."""
+    return np.dot(image[:, :, 0:3], [0.299, 0.587, 0.114]).astype(np.uint8)
+
+
+def binarize(image, threshold_value):
+    """Binarizes a grayscale image based on a given threshold value, setting values to 1 or 0 accordingly."""
+    binarized = np.where(image > threshold_value, 1, 0)
+
+    return binarized
+
+
+def transform(image, kind, kernel=np.ones((3, 3))):
+    """Simple image transformation using one of two available filter functions: Erosion and Dilation.
+
+    Args:
+        image:
+        kind: Can be either 'erosion', in which case the :func:np.max function is called,
+        or 'dilation', when :func:np.min is used instead.
+        kernel: n x n kernel with shape < :attr:image.shape, to be used when applying convolution to original image
+
+    Returns:
+
+    """
+    if kind == "erosion":
+        constant = 1
+        apply = np.max
+    else:
+        constant = 0
+        apply = np.min
+
+    center_x, center_y = (x // 2 for x in kernel.shape)
+
+    # Use padded image when applying convolotion to not go out of bounds of the original the image
+    transformed = np.zeros(image.shape, dtype=np.uint8)
+    padded = np.pad(image, 1, 'constant', constant_values=constant)
+
+    for x in range(center_x, padded.shape[0] - center_x):
+        for y in range(center_y, padded.shape[1] - center_y):
+
+            center = padded[x - center_x: x + center_x + 1, y - center_y: y + center_y + 1]
+            # Apply transformation method to the centered section of the image
+            transformed[x - center_x, y - center_y] = apply(center[kernel == 1])
+
+    return transformed
+
+
+def opening_filter(image, kernel=np.ones((3, 3))):
+    """Opening filter, defined as the sequence of erosion and then a dilation filter on the same image."""
+    return transform(
+        transform(image, 'dilation', kernel),
+        'erosion', kernel
+    )
+
+
+def closing_filter(image, kernel=np.ones((3, 3))):
+    """Opening filter, defined as the sequence of dilation and then erosion filter on the same image."""
+    return transform(
+        transform(image, 'erosion', kernel),
+        'dilation', kernel
+    )
+
+
+def binary_mask(image_gray, image_map):
+    """Apply binary mask, or thresholding based on bit mask value (mapping mask is 1 or 0)."""
+    true_mask, false_mask = np.array(image_gray, copy=True), np.array(image_gray, copy=True)
+    true_mask[image_map == 1] = 1
+    false_mask[image_map == 0] = 0
+
+    return true_mask, false_mask
+
+
+def matrix_concurrency(image, coordinate):
+    """Calculate sample co-occurrence matrix based on input image as well as selected coordinates on image.
+    Implementation is made using basic iteration, as function to be performed (np.max) is non-linear and therefore
+    not usable on the Fourier Transform domain."""
+    matrix = np.zeros([np.max(image) + 1, np.max(image) + 1])
+
+    offset_x, offset_y = coordinate[0], coordinate[1]
+
+    for x in range(1, image.shape[0] - 1):
+        for y in range(1, image.shape[1] - 1):
+
+            base_pixel = image[x, y]
+            offset_pixel = image[x + offset_x, y + offset_y]
+
+            matrix[base_pixel, offset_pixel] += 1
+
+    return matrix / np.sum(matrix)
+
+
+def haralick_descriptors(matrix):
+    """Calculates all 8 Haralick descriptors based on co-occurence input matrix.
+    All descriptors are as follows:
+    Maximum probability, Inverse Difference, Homogeneity, Entropy, Energy, Dissimilarity, Contrast and Correlation
+
+    Args:
+        matrix: Co-occurence matrix to use as base for calculating descriptors.
+
+    Returns:
+        Reverse ordered list of resulting descriptors
+    """
+    # Function np.indices could be used for bigger input types, but np.ogrid works just fine
+    i, j = np.ogrid[0:matrix.shape[0], 0:matrix.shape[1]]  # np.indices()
+
+    # Pre-calculate frequent multiplication and subtraction
+    prod = np.multiply(i, j)
+    sub = np.subtract(i, j)
+
+    # Calculate numerical value of Maximum Probability
+    maximum_prob = np.max(matrix)
+    # Using the definition for each descriptor individually to calculate its matrix
+    correlation = prod * matrix
+    energy = np.power(matrix, 2)
+    contrast = matrix * np.power(sub, 2)
+
+    dissimilarity = matrix * np.abs(sub)
+    inverse_difference = matrix / (1 + np.abs(sub))
+    homogeneity = matrix / (1 + np.power(sub, 2))
+    entropy = -(matrix[matrix > 0] * np.log(matrix[matrix > 0]))
+
+    # Sum values for descriptors ranging from the first one to the last, as all are their respective origin matrix
+    # and not the resulting value yet.
+    descriptors = [maximum_prob, correlation.sum(), energy.sum(), contrast.sum(),
+                   dissimilarity.sum(),  inverse_difference.sum(), homogeneity.sum(), entropy.sum()]
+    return descriptors
+
+
+def get_descriptors(masks, coordinate):
+    """Calculate all Haralick descriptors for a sequence of different co-occurrence matrices, given input
+    masks and coordinates."""
+
+    descriptors = list()
+    for mask in masks:
+        descriptors.append(haralick_descriptors(
+            matrix_concurrency(mask, coordinate))
+        )
+    # Concatenate each individual descriptor into one single list containing sequence of descriptors
+    return np.concatenate(descriptors, axis=None)
+
+
+def euclidean(point_1: np.ndarray, point_2: np.ndarray):
+    """Simple method for calculating the euclidean distance between two points, with type np.ndarray."""
+    return np.sqrt(np.sum(np.square(point_1 - point_2)))
+
+
+def get_distances(descriptors, base):
+    """Calculate all Euclidian distances between a selected base descriptor and all other Haralick descriptors
+    The resulting comparison is return in decreasing order, showing which descriptor is the most similar to the
+    selected base.
+
+    Args:
+        descriptors: Haralick descriptors to compare with base index
+        base: Haralick descriptor index to use as base when calculating respective euclidean distance to other descriptors.
+
+    Returns:
+        Ordered distances between descriptors
+
+    """
+    distances = []
+
+    for description in descriptors:
+        distances.append(euclidean(description, descriptors[base]))
+    # Normalize distances between range [0, 1]
+    distances = normalize_array(distances, 1)
+    return sorted(enumerate(distances), key=lambda tup: tup[1])
+
+
+def main():
+    # Index to compare haralick descriptors to
+    index = int(input())
+    q_value = [int(value) for value in input().split()]
+
+    # Format is the respective filter to apply, can be either 1 for the opening filter or else for the closing
+    parameters = {'format': int(input()), 'threshold': int(input())}
+
+    # Number of images to perform methods on
+    b_number = int(input())
+
+    files, descriptors = (list(), list())
+
+    for _ in range(b_number):
+        file = input().rstrip()
+        files.append(file)
+
+        # Open given image and calculate morphological filter, respective masks and correspondent Harralick Descriptors.
+        image = imageio.imread(file).astype(np.float32)
+        gray = grayscale(image)
+        threshold = binarize(gray, parameters['threshold'])
+
+        morphological = opening_filter(threshold) if parameters['format'] == 1 else closing_filter(threshold)
+        masks = binary_mask(gray, morphological)
+        descriptors.append(get_descriptors(masks, q_value))
+
+    # Transform ordered distances array into a sequence of indexes corresponding to original file position
+    distances = get_distances(descriptors, index)
+    indexed_distances = np.array(distances).astype(np.uint8)[:, 0]
+
+    # Finally, print distances considering the Haralick descriptions from the base file to
+    # all other images using the morphology method of choice.
+    print(f"Query: {files[index]}")
+    print("Ranking:")
+    for idx, file_idx in enumerate(indexed_distances):
+        print(f"({idx}) {files[file_idx]}", end="\n")
+
+
+if __name__ == "__main__":
+    main()

From 88b014f7ab6ee9d8b50c4e6b6a02e38d3a8e7d52 Mon Sep 17 00:00:00 2001
From: Rafael Zimmer <rzimmerdev@monmo.localdomain>
Date: Wed, 19 Oct 2022 01:25:19 -0300
Subject: [PATCH 02/17] Working on creating Unit Testing for Haralick
 Descriptors module

---
 ...ck_descriptors => haralick_descriptors.py} | 199 ++++++++++++------
 1 file changed, 131 insertions(+), 68 deletions(-)
 rename computer_vision/{haralick_descriptors => haralick_descriptors.py} (52%)

diff --git a/computer_vision/haralick_descriptors b/computer_vision/haralick_descriptors.py
similarity index 52%
rename from computer_vision/haralick_descriptors
rename to computer_vision/haralick_descriptors.py
index 6f8d6bd66b80..1006f4db9435 100644
--- a/computer_vision/haralick_descriptors
+++ b/computer_vision/haralick_descriptors.py
@@ -1,16 +1,36 @@
+import imageio.v2 as imageio
+import numpy as np
+
+
 def rmse(original: np.ndarray, reference: np.ndarray) -> float:
-    """Simple implementation of Root Mean Squared Error for two N dimensional numpy arrays."""
+    """Simple implementation of Root Mean Squared Error
+    for two N dimensional numpy arrays.
+    >>> rmse(np.array([1, 2, 3]), np.array([1, 2, 3]))
+    0.0
+    >>> rmse(np.array([1, 2, 3]), np.array([2, 2, 2]))
+    1.0
+    >>> rmse(np.array([1, 2, 3]), np.array([6, 4, 2]))
+    30.0
+    """
     return np.sqrt(((original - reference) ** 2).mean())
 
 
 def normalize_image(image: np.ndarray, cap: float = 255) -> np.ndarray:
-    """Normalizes image in Numpy 2D array format, between ranges 0-cap, as to fit uint8 type.
+    """
+    Normalizes image in Numpy 2D array format, between ranges 0-cap,
+    as to fit uint8 type.
 
     Args:
         image: 2D numpy array representing image as matrix, with values in any range
         cap: Maximum cap amount for normalization
     Returns:
         return 2D numpy array of type uint8, corresponding to limited range matrix
+
+    >>> normalize_image(np.array([[1, 2, 3], [4, 5, 10]]), cap=1)
+    array([[ 0,         0.111111,  0.222222],
+           [ 0.333333,  0.444444,  1.      ]], dtype=float32)
+    >>> normalize_image(np.array([[4, 4, 3], [1, 7, 2]]))
+
     """
     normalized = (image - np.min(image)) / (np.max(image) - np.min(image)) * cap
     return normalized.astype(np.uint8)
@@ -29,11 +49,6 @@ def normalize_array(array, cap: float = 1):
     return normalized
 
 
-def euclidean(point_1: np.ndarray, point_2: np.ndarray):
-    """Simple method for calculating the euclidean distance between two points, with type np.ndarray."""
-    return np.sqrt(np.sum(np.square(point_1 - point_2)))
-
-
 def grayscale(image: np.ndarray) -> np.ndarray:
     """Uses luminance weights to transform RGB channel to greyscale, by
     taking the dot product between the channel and the weights."""
@@ -41,24 +56,33 @@ def grayscale(image: np.ndarray) -> np.ndarray:
 
 
 def binarize(image, threshold_value):
-    """Binarizes a grayscale image based on a given threshold value, setting values to 1 or 0 accordingly."""
+    """
+    Binarizes a grayscale image based on a given threshold value,
+    setting values to 1 or 0 accordingly.
+    """
     binarized = np.where(image > threshold_value, 1, 0)
 
     return binarized
 
 
-def transform(image, kind, kernel=np.ones((3, 3))):
-    """Simple image transformation using one of two available filter functions: Erosion and Dilation.
+def transform(image, kind, kernel=None):
+    """
+    Simple image transformation using one of two available filter functions:
+    Erosion and Dilation.
 
     Args:
         image:
-        kind: Can be either 'erosion', in which case the :func:np.max function is called,
-        or 'dilation', when :func:np.min is used instead.
-        kernel: n x n kernel with shape < :attr:image.shape, to be used when applying convolution to original image
+        kind: Can be either 'erosion', in which case the :func:np.max
+              function is called, or 'dilation', when :func:np.min is used instead.
+        kernel: n x n kernel with shape < :attr:image.shape,
+              to be used when applying convolution to original image
 
     Returns:
 
     """
+    if kernel is None:
+        kernel = np.ones((3, 3))
+
     if kind == "erosion":
         constant = 1
         apply = np.max
@@ -68,39 +92,52 @@ def transform(image, kind, kernel=np.ones((3, 3))):
 
     center_x, center_y = (x // 2 for x in kernel.shape)
 
-    # Use padded image when applying convolotion to not go out of bounds of the original the image
+    # Use padded image when applying convolotion
+    # to not go out of bounds of the original the image
     transformed = np.zeros(image.shape, dtype=np.uint8)
-    padded = np.pad(image, 1, 'constant', constant_values=constant)
+    padded = np.pad(image, 1, "constant", constant_values=constant)
 
     for x in range(center_x, padded.shape[0] - center_x):
         for y in range(center_y, padded.shape[1] - center_y):
-
-            center = padded[x - center_x: x + center_x + 1, y - center_y: y + center_y + 1]
+            center = padded[
+                x - center_x : x + center_x + 1, y - center_y : y + center_y + 1
+            ]
             # Apply transformation method to the centered section of the image
             transformed[x - center_x, y - center_y] = apply(center[kernel == 1])
 
     return transformed
 
 
-def opening_filter(image, kernel=np.ones((3, 3))):
-    """Opening filter, defined as the sequence of erosion and then a dilation filter on the same image."""
-    return transform(
-        transform(image, 'dilation', kernel),
-        'erosion', kernel
-    )
+def opening_filter(image, kernel=None):
+    """
+    Opening filter, defined as the sequence of
+    erosion and then a dilation filter on the same image.
+    """
+    if kernel is None:
+        np.ones((3, 3))
 
+    return transform(transform(image, "dilation", kernel), "erosion", kernel)
 
-def closing_filter(image, kernel=np.ones((3, 3))):
-    """Opening filter, defined as the sequence of dilation and then erosion filter on the same image."""
-    return transform(
-        transform(image, 'erosion', kernel),
-        'dilation', kernel
-    )
+
+def closing_filter(image, kernel=None):
+    """
+    Opening filter, defined as the sequence of
+    dilation and then erosion filter on the same image.
+    """
+    if kernel is None:
+        np.ones((3, 3))
+
+    return transform(transform(image, "erosion", kernel), "dilation", kernel)
 
 
 def binary_mask(image_gray, image_map):
-    """Apply binary mask, or thresholding based on bit mask value (mapping mask is 1 or 0)."""
-    true_mask, false_mask = np.array(image_gray, copy=True), np.array(image_gray, copy=True)
+    """
+    Apply binary mask, or thresholding based
+    on bit mask value (mapping mask is 1 or 0).
+    """
+    true_mask, false_mask = np.array(image_gray, copy=True), np.array(
+        image_gray, copy=True
+    )
     true_mask[image_map == 1] = 1
     false_mask[image_map == 0] = 0
 
@@ -108,16 +145,19 @@ def binary_mask(image_gray, image_map):
 
 
 def matrix_concurrency(image, coordinate):
-    """Calculate sample co-occurrence matrix based on input image as well as selected coordinates on image.
-    Implementation is made using basic iteration, as function to be performed (np.max) is non-linear and therefore
-    not usable on the Fourier Transform domain."""
+    """
+    Calculate sample co-occurrence matrix based on input image
+    as well as selected coordinates on image.
+    Implementation is made using basic iteration,
+    as function to be performed (np.max) is non-linear and therefore
+    not usable on the Fourier Transform domain.
+    """
     matrix = np.zeros([np.max(image) + 1, np.max(image) + 1])
 
     offset_x, offset_y = coordinate[0], coordinate[1]
 
     for x in range(1, image.shape[0] - 1):
         for y in range(1, image.shape[1] - 1):
-
             base_pixel = image[x, y]
             offset_pixel = image[x + offset_x, y + offset_y]
 
@@ -129,7 +169,8 @@ def matrix_concurrency(image, coordinate):
 def haralick_descriptors(matrix):
     """Calculates all 8 Haralick descriptors based on co-occurence input matrix.
     All descriptors are as follows:
-    Maximum probability, Inverse Difference, Homogeneity, Entropy, Energy, Dissimilarity, Contrast and Correlation
+    Maximum probability, Inverse Difference, Homogeneity, Entropy,
+    Energy, Dissimilarity, Contrast and Correlation
 
     Args:
         matrix: Co-occurence matrix to use as base for calculating descriptors.
@@ -137,8 +178,9 @@ def haralick_descriptors(matrix):
     Returns:
         Reverse ordered list of resulting descriptors
     """
-    # Function np.indices could be used for bigger input types, but np.ogrid works just fine
-    i, j = np.ogrid[0:matrix.shape[0], 0:matrix.shape[1]]  # np.indices()
+    # Function np.indices could be used for bigger input types,
+    # but np.ogrid works just fine
+    i, j = np.ogrid[0 : matrix.shape[0], 0 : matrix.shape[1]]  # np.indices()
 
     # Pre-calculate frequent multiplication and subtraction
     prod = np.multiply(i, j)
@@ -156,48 +198,62 @@ def haralick_descriptors(matrix):
     homogeneity = matrix / (1 + np.power(sub, 2))
     entropy = -(matrix[matrix > 0] * np.log(matrix[matrix > 0]))
 
-    # Sum values for descriptors ranging from the first one to the last, as all are their respective origin matrix
-    # and not the resulting value yet.
-    descriptors = [maximum_prob, correlation.sum(), energy.sum(), contrast.sum(),
-                   dissimilarity.sum(),  inverse_difference.sum(), homogeneity.sum(), entropy.sum()]
+    # Sum values for descriptors ranging from the first one to the last,
+    # as all are their respective origin matrix and not the resulting value yet.
+    descriptors = [
+        maximum_prob,
+        correlation.sum(),
+        energy.sum(),
+        contrast.sum(),
+        dissimilarity.sum(),
+        inverse_difference.sum(),
+        homogeneity.sum(),
+        entropy.sum(),
+    ]
     return descriptors
 
 
 def get_descriptors(masks, coordinate):
-    """Calculate all Haralick descriptors for a sequence of different co-occurrence matrices, given input
-    masks and coordinates."""
+    """
+    Calculate all Haralick descriptors for a sequence of
+    different co-occurrence matrices, given input masks and coordinates.
+    """
+    descriptors = np.zeros((len(masks), 8))
+    for idx, mask in enumerate(masks):
+        descriptors[idx] = haralick_descriptors(matrix_concurrency(mask, coordinate))
 
-    descriptors = list()
-    for mask in masks:
-        descriptors.append(haralick_descriptors(
-            matrix_concurrency(mask, coordinate))
-        )
-    # Concatenate each individual descriptor into one single list containing sequence of descriptors
+    # Concatenate each individual descriptor into
+    # one single list containing sequence of descriptors
     return np.concatenate(descriptors, axis=None)
 
 
 def euclidean(point_1: np.ndarray, point_2: np.ndarray):
-    """Simple method for calculating the euclidean distance between two points, with type np.ndarray."""
+    """
+    Simple method for calculating the euclidean distance between two points,
+    with type np.ndarray.
+    """
     return np.sqrt(np.sum(np.square(point_1 - point_2)))
 
 
 def get_distances(descriptors, base):
-    """Calculate all Euclidian distances between a selected base descriptor and all other Haralick descriptors
-    The resulting comparison is return in decreasing order, showing which descriptor is the most similar to the
-    selected base.
+    """
+    Calculate all Euclidean distances between a selected base descriptor
+    and all other Haralick descriptors
+    The resulting comparison is return in decreasing order,
+    showing which descriptor is the most similar to the selected base.
 
     Args:
         descriptors: Haralick descriptors to compare with base index
-        base: Haralick descriptor index to use as base when calculating respective euclidean distance to other descriptors.
+        base: Haralick descriptor index to use as base when calculating respective
+        euclidean distance to other descriptors.
 
     Returns:
         Ordered distances between descriptors
-
     """
-    distances = []
+    distances = np.zeros(descriptors.shape[0])
 
-    for description in descriptors:
-        distances.append(euclidean(description, descriptors[base]))
+    for idx, description in enumerate(descriptors):
+        distances[idx] = euclidean(description, descriptors[base])
     # Normalize distances between range [0, 1]
     distances = normalize_array(distances, 1)
     return sorted(enumerate(distances), key=lambda tup: tup[1])
@@ -208,33 +264,40 @@ def main():
     index = int(input())
     q_value = [int(value) for value in input().split()]
 
-    # Format is the respective filter to apply, can be either 1 for the opening filter or else for the closing
-    parameters = {'format': int(input()), 'threshold': int(input())}
+    # Format is the respective filter to apply,
+    # can be either 1 for the opening filter or else for the closing
+    parameters = {"format": int(input()), "threshold": int(input())}
 
     # Number of images to perform methods on
     b_number = int(input())
 
-    files, descriptors = (list(), list())
+    files, descriptors = (np.array([]), np.array([]))
 
     for _ in range(b_number):
         file = input().rstrip()
         files.append(file)
 
-        # Open given image and calculate morphological filter, respective masks and correspondent Harralick Descriptors.
+        # Open given image and calculate morphological filter,
+        # respective masks and correspondent Harralick Descriptors.
         image = imageio.imread(file).astype(np.float32)
         gray = grayscale(image)
-        threshold = binarize(gray, parameters['threshold'])
+        threshold = binarize(gray, parameters["threshold"])
 
-        morphological = opening_filter(threshold) if parameters['format'] == 1 else closing_filter(threshold)
+        morphological = (
+            opening_filter(threshold)
+            if parameters["format"] == 1
+            else closing_filter(threshold)
+        )
         masks = binary_mask(gray, morphological)
         descriptors.append(get_descriptors(masks, q_value))
 
-    # Transform ordered distances array into a sequence of indexes corresponding to original file position
+    # Transform ordered distances array into a sequence of indexes
+    # corresponding to original file position
     distances = get_distances(descriptors, index)
     indexed_distances = np.array(distances).astype(np.uint8)[:, 0]
 
-    # Finally, print distances considering the Haralick descriptions from the base file to
-    # all other images using the morphology method of choice.
+    # Finally, print distances considering the Haralick descriptions from the base
+    # file to all other images using the morphology method of choice.
     print(f"Query: {files[index]}")
     print("Ranking:")
     for idx, file_idx in enumerate(indexed_distances):

From bc6a1894f84173910e95ba8544ca64f42788970c Mon Sep 17 00:00:00 2001
From: Rafael Zimmer <rzimmerdev@monmo.localdomain>
Date: Wed, 19 Oct 2022 01:59:19 -0300
Subject: [PATCH 03/17] Type hinting for Haralick descriptors

---
 computer_vision/haralick_descriptors.py | 22 +++++++++++-----------
 1 file changed, 11 insertions(+), 11 deletions(-)

diff --git a/computer_vision/haralick_descriptors.py b/computer_vision/haralick_descriptors.py
index 1006f4db9435..6c64632f8cf1 100644
--- a/computer_vision/haralick_descriptors.py
+++ b/computer_vision/haralick_descriptors.py
@@ -36,7 +36,7 @@ def normalize_image(image: np.ndarray, cap: float = 255) -> np.ndarray:
     return normalized.astype(np.uint8)
 
 
-def normalize_array(array, cap: float = 1):
+def normalize_array(array: np.ndarray, cap: float = 1) -> np.ndarray:
     """Normalizes a 1D array, between ranges 0-cap.
 
     Args:
@@ -55,7 +55,7 @@ def grayscale(image: np.ndarray) -> np.ndarray:
     return np.dot(image[:, :, 0:3], [0.299, 0.587, 0.114]).astype(np.uint8)
 
 
-def binarize(image, threshold_value):
+def binarize(image: np.ndarray, threshold_value) -> np.ndarray:
     """
     Binarizes a grayscale image based on a given threshold value,
     setting values to 1 or 0 accordingly.
@@ -65,7 +65,7 @@ def binarize(image, threshold_value):
     return binarized
 
 
-def transform(image, kind, kernel=None):
+def transform(image: np.ndarray, kind, kernel=None) -> np.ndarray:
     """
     Simple image transformation using one of two available filter functions:
     Erosion and Dilation.
@@ -108,7 +108,7 @@ def transform(image, kind, kernel=None):
     return transformed
 
 
-def opening_filter(image, kernel=None):
+def opening_filter(image: np.ndarray, kernel: np.ndarray = None) -> np.ndarray:
     """
     Opening filter, defined as the sequence of
     erosion and then a dilation filter on the same image.
@@ -119,7 +119,7 @@ def opening_filter(image, kernel=None):
     return transform(transform(image, "dilation", kernel), "erosion", kernel)
 
 
-def closing_filter(image, kernel=None):
+def closing_filter(image: np.ndarray, kernel: np.ndarray = None) -> np.ndarray:
     """
     Opening filter, defined as the sequence of
     dilation and then erosion filter on the same image.
@@ -130,7 +130,7 @@ def closing_filter(image, kernel=None):
     return transform(transform(image, "erosion", kernel), "dilation", kernel)
 
 
-def binary_mask(image_gray, image_map):
+def binary_mask(image_gray: np.ndarray, image_map: np.ndarray) -> np.ndarray:
     """
     Apply binary mask, or thresholding based
     on bit mask value (mapping mask is 1 or 0).
@@ -144,7 +144,7 @@ def binary_mask(image_gray, image_map):
     return true_mask, false_mask
 
 
-def matrix_concurrency(image, coordinate):
+def matrix_concurrency(image: np.ndarray, coordinate) -> np.ndarray:
     """
     Calculate sample co-occurrence matrix based on input image
     as well as selected coordinates on image.
@@ -166,7 +166,7 @@ def matrix_concurrency(image, coordinate):
     return matrix / np.sum(matrix)
 
 
-def haralick_descriptors(matrix):
+def haralick_descriptors(matrix: np.ndarray) -> list:
     """Calculates all 8 Haralick descriptors based on co-occurence input matrix.
     All descriptors are as follows:
     Maximum probability, Inverse Difference, Homogeneity, Entropy,
@@ -213,7 +213,7 @@ def haralick_descriptors(matrix):
     return descriptors
 
 
-def get_descriptors(masks, coordinate):
+def get_descriptors(masks, coordinate) -> np.ndarray:
     """
     Calculate all Haralick descriptors for a sequence of
     different co-occurrence matrices, given input masks and coordinates.
@@ -227,7 +227,7 @@ def get_descriptors(masks, coordinate):
     return np.concatenate(descriptors, axis=None)
 
 
-def euclidean(point_1: np.ndarray, point_2: np.ndarray):
+def euclidean(point_1: np.ndarray, point_2: np.ndarray) -> np.float32:
     """
     Simple method for calculating the euclidean distance between two points,
     with type np.ndarray.
@@ -235,7 +235,7 @@ def euclidean(point_1: np.ndarray, point_2: np.ndarray):
     return np.sqrt(np.sum(np.square(point_1 - point_2)))
 
 
-def get_distances(descriptors, base):
+def get_distances(descriptors, base) -> np.ndarray:
     """
     Calculate all Euclidean distances between a selected base descriptor
     and all other Haralick descriptors

From 6e23cc094f661b5ec21a1e9d9e4cfef2c5c0e8b8 Mon Sep 17 00:00:00 2001
From: Rafael Zimmer <rzimmerdev@gmail.com>
Date: Sun, 27 Nov 2022 13:44:28 -0300
Subject: [PATCH 04/17] Fixed docstrings, unit testing and formatting choices

---
 computer_vision/haralick_descriptors.py | 160 ++++++++++++++++++------
 1 file changed, 119 insertions(+), 41 deletions(-)

diff --git a/computer_vision/haralick_descriptors.py b/computer_vision/haralick_descriptors.py
index 6c64632f8cf1..8cf0d702ce32 100644
--- a/computer_vision/haralick_descriptors.py
+++ b/computer_vision/haralick_descriptors.py
@@ -1,21 +1,29 @@
+"""
+https://en.wikipedia.org/wiki/Image_texture
+https://en.wikipedia.org/wiki/Co-occurrence_matrix#Application_to_image_analysis
+"""
+from typing import Any, Union
+
 import imageio.v2 as imageio
 import numpy as np
 
 
-def rmse(original: np.ndarray, reference: np.ndarray) -> float:
+def root_mean_square_error(original: np.ndarray, reference: np.ndarray) -> float:
     """Simple implementation of Root Mean Squared Error
     for two N dimensional numpy arrays.
-    >>> rmse(np.array([1, 2, 3]), np.array([1, 2, 3]))
-    0.0
-    >>> rmse(np.array([1, 2, 3]), np.array([2, 2, 2]))
-    1.0
-    >>> rmse(np.array([1, 2, 3]), np.array([6, 4, 2]))
-    30.0
+
+    Examples:
+        >>> root_mean_square_error(np.array([1, 2, 3]), np.array([1, 2, 3]))
+        0.0
+        >>> root_mean_square_error(np.array([1, 2, 3]), np.array([2, 2, 2]))
+        0.816496580927726
+        >>> root_mean_square_error(np.array([1, 2, 3]), np.array([6, 4, 2]))
+        3.1622776601683795
     """
     return np.sqrt(((original - reference) ** 2).mean())
 
 
-def normalize_image(image: np.ndarray, cap: float = 255) -> np.ndarray:
+def normalize_image(image: np.ndarray, cap: float = 255.0, data_type=np.uint8) -> np.ndarray:
     """
     Normalizes image in Numpy 2D array format, between ranges 0-cap,
     as to fit uint8 type.
@@ -23,17 +31,20 @@ def normalize_image(image: np.ndarray, cap: float = 255) -> np.ndarray:
     Args:
         image: 2D numpy array representing image as matrix, with values in any range
         cap: Maximum cap amount for normalization
+        data_type: numpy data type to set output variable to
     Returns:
         return 2D numpy array of type uint8, corresponding to limited range matrix
 
-    >>> normalize_image(np.array([[1, 2, 3], [4, 5, 10]]), cap=1)
-    array([[ 0,         0.111111,  0.222222],
-           [ 0.333333,  0.444444,  1.      ]], dtype=float32)
-    >>> normalize_image(np.array([[4, 4, 3], [1, 7, 2]]))
-
+    Examples:
+        >>> normalize_image(np.array([[1, 2, 3], [4, 5, 10]]), cap=1.0, data_type=np.float64)
+        array([[0.        , 0.11111111, 0.22222222],
+               [0.33333333, 0.44444444, 1.        ]])
+        >>> normalize_image(np.array([[4, 4, 3], [1, 7, 2]]))
+        array([[127, 127,  85],
+               [  0, 255,  42]], dtype=uint8)
     """
     normalized = (image - np.min(image)) / (np.max(image) - np.min(image)) * cap
-    return normalized.astype(np.uint8)
+    return normalized.astype(data_type)
 
 
 def normalize_array(array: np.ndarray, cap: float = 1) -> np.ndarray:
@@ -43,44 +54,75 @@ def normalize_array(array: np.ndarray, cap: float = 1) -> np.ndarray:
         array: List containing values to be normalized between cap range.
         cap: Maximum cap amount for normalization.
     Returns:
-        return 1D numpy array , corresponding to limited range array
+        return 1D numpy array, corresponding to limited range array
+
+    Examples:
+        >>> normalize_array(np.array([2, 3, 5, 7]))
+        array([0. , 0.2, 0.6, 1. ])
+        >>> normalize_array(np.array([[5], [7], [11], [13]]))
+        array([[0.  ],
+               [0.25],
+               [0.75],
+               [1.  ]])
     """
-    normalized = (array - np.min(array)) / (np.max(array) - np.min(array)) * cap
-    return normalized
+    return (array - np.min(array)) / (np.max(array) - np.min(array)) * cap
 
 
 def grayscale(image: np.ndarray) -> np.ndarray:
-    """Uses luminance weights to transform RGB channel to greyscale, by
-    taking the dot product between the channel and the weights."""
+    """
+    Uses luminance weights to transform RGB channel to greyscale, by
+    taking the dot product between the channel and the weights.
+
+    Example:
+        >>> grayscale(np.array([[[108, 201, 72], [255, 11,  127]], [[56,  56,  56], [128, 255, 107]]]))
+        array([[158,  97],
+               [ 56, 200]], dtype=uint8)
+    """
     return np.dot(image[:, :, 0:3], [0.299, 0.587, 0.114]).astype(np.uint8)
 
 
-def binarize(image: np.ndarray, threshold_value) -> np.ndarray:
+def binarize(image: np.ndarray, threshold: float = 127.0) -> np.ndarray:
     """
     Binarizes a grayscale image based on a given threshold value,
     setting values to 1 or 0 accordingly.
-    """
-    binarized = np.where(image > threshold_value, 1, 0)
 
-    return binarized
+    Examples:
+        >>> binarize(np.array([[128, 255], [101, 156]]))
+        array([[1, 1],
+               [0, 1]])
+        >>> binarize(np.array([[0.07, 1], [0.51, 0.3]]), threshold=0.5)
+        array([[0, 1],
+               [1, 0]])
+    """
+    return np.where(image > threshold, 1, 0)
 
 
-def transform(image: np.ndarray, kind, kernel=None) -> np.ndarray:
+def transform(image: np.ndarray, kind: str, kernel=None) -> np.ndarray:
     """
     Simple image transformation using one of two available filter functions:
     Erosion and Dilation.
 
     Args:
-        image:
+        image: binarized input image, onto which to apply transformation
         kind: Can be either 'erosion', in which case the :func:np.max
               function is called, or 'dilation', when :func:np.min is used instead.
         kernel: n x n kernel with shape < :attr:image.shape,
               to be used when applying convolution to original image
 
     Returns:
-
+        returns a numpy array with same shape as input image, corresponding to applied binary transformation.
+
+    Examples:
+        >>> img = np.array([[1, 0.5], [0.2, 0.7]])
+        >>> img = binarize(img, threshold=0.5)
+        >>> transform(img, 'erosion')
+        array([[1, 1],
+               [1, 1]], dtype=uint8)
+        >>> transform(img, 'dilation')
+        array([[0, 0],
+               [0, 0]], dtype=uint8)
     """
-    if kernel is None:
+    if not kernel:
         kernel = np.ones((3, 3))
 
     if kind == "erosion":
@@ -100,8 +142,8 @@ def transform(image: np.ndarray, kind, kernel=None) -> np.ndarray:
     for x in range(center_x, padded.shape[0] - center_x):
         for y in range(center_y, padded.shape[1] - center_y):
             center = padded[
-                x - center_x : x + center_x + 1, y - center_y : y + center_y + 1
-            ]
+                     x - center_x: x + center_x + 1, y - center_y: y + center_y + 1
+                     ]
             # Apply transformation method to the centered section of the image
             transformed[x - center_x, y - center_y] = apply(center[kernel == 1])
 
@@ -112,8 +154,15 @@ def opening_filter(image: np.ndarray, kernel: np.ndarray = None) -> np.ndarray:
     """
     Opening filter, defined as the sequence of
     erosion and then a dilation filter on the same image.
+
+    Examples:
+        >>> img = np.array([[1, 0.5], [0.2, 0.7]])
+        >>> img = binarize(img, threshold=0.5)
+        >>> opening_filter(img)
+        array([[1, 1],
+               [1, 1]], dtype=uint8)
     """
-    if kernel is None:
+    if not kernel:
         np.ones((3, 3))
 
     return transform(transform(image, "dilation", kernel), "erosion", kernel)
@@ -123,6 +172,13 @@ def closing_filter(image: np.ndarray, kernel: np.ndarray = None) -> np.ndarray:
     """
     Opening filter, defined as the sequence of
     dilation and then erosion filter on the same image.
+
+    Examples:
+        >>> img = np.array([[1, 0.5], [0.2, 0.7]])
+        >>> img = binarize(img, threshold=0.5)
+        >>> closing_filter(img)
+        array([[0, 0],
+               [0, 0]], dtype=uint8)
     """
     if kernel is None:
         np.ones((3, 3))
@@ -130,10 +186,24 @@ def closing_filter(image: np.ndarray, kernel: np.ndarray = None) -> np.ndarray:
     return transform(transform(image, "erosion", kernel), "dilation", kernel)
 
 
-def binary_mask(image_gray: np.ndarray, image_map: np.ndarray) -> np.ndarray:
+def binary_mask(
+        image_gray: np.ndarray, image_map: np.ndarray
+) -> tuple[np.ndarray, np.ndarray]:
     """
     Apply binary mask, or thresholding based
-    on bit mask value (mapping mask is 1 or 0).
+    on bit mask value (mapping mask is binary).
+
+    Returns the mapped true value mask and its complementary false value mask.
+
+    Example:
+        >>> img = np.array([[[108, 201, 72], [255, 11,  127]], [[56,  56,  56], [128, 255, 107]]])
+        >>> gray = grayscale(img)
+        >>> binary = binarize(gray)
+        >>> morphological = opening_filter(binary)
+        >>> binary_mask(gray, morphological)
+        (array([[1, 1],
+               [1, 1]], dtype=uint8), array([[158,  97],
+               [ 56, 200]], dtype=uint8))
     """
     true_mask, false_mask = np.array(image_gray, copy=True), np.array(
         image_gray, copy=True
@@ -144,13 +214,16 @@ def binary_mask(image_gray: np.ndarray, image_map: np.ndarray) -> np.ndarray:
     return true_mask, false_mask
 
 
-def matrix_concurrency(image: np.ndarray, coordinate) -> np.ndarray:
+def matrix_concurrency(
+        image: np.ndarray, coordinate
+) -> np.ndarray:
     """
     Calculate sample co-occurrence matrix based on input image
     as well as selected coordinates on image.
+
     Implementation is made using basic iteration,
     as function to be performed (np.max) is non-linear and therefore
-    not usable on the Fourier Transform domain.
+    not callable on the frequency domain.
     """
     matrix = np.zeros([np.max(image) + 1, np.max(image) + 1])
 
@@ -180,7 +253,7 @@ def haralick_descriptors(matrix: np.ndarray) -> list:
     """
     # Function np.indices could be used for bigger input types,
     # but np.ogrid works just fine
-    i, j = np.ogrid[0 : matrix.shape[0], 0 : matrix.shape[1]]  # np.indices()
+    i, j = np.ogrid[0: matrix.shape[0], 0: matrix.shape[1]]  # np.indices()
 
     # Pre-calculate frequent multiplication and subtraction
     prod = np.multiply(i, j)
@@ -200,7 +273,7 @@ def haralick_descriptors(matrix: np.ndarray) -> list:
 
     # Sum values for descriptors ranging from the first one to the last,
     # as all are their respective origin matrix and not the resulting value yet.
-    descriptors = [
+    return [
         maximum_prob,
         correlation.sum(),
         energy.sum(),
@@ -210,10 +283,9 @@ def haralick_descriptors(matrix: np.ndarray) -> list:
         homogeneity.sum(),
         entropy.sum(),
     ]
-    return descriptors
 
 
-def get_descriptors(masks, coordinate) -> np.ndarray:
+def get_descriptors(masks: tuple[np.ndarray, np.ndarray], coordinate) -> np.ndarray:
     """
     Calculate all Haralick descriptors for a sequence of
     different co-occurrence matrices, given input masks and coordinates.
@@ -231,11 +303,17 @@ def euclidean(point_1: np.ndarray, point_2: np.ndarray) -> np.float32:
     """
     Simple method for calculating the euclidean distance between two points,
     with type np.ndarray.
+
+    Example:
+        >>> a = np.array([1, 0, -2])
+        >>> b = np.array([2, -1, 1])
+        >>> euclidean(a, b)
+        3.3166247903554
     """
     return np.sqrt(np.sum(np.square(point_1 - point_2)))
 
 
-def get_distances(descriptors, base) -> np.ndarray:
+def get_distances(descriptors, base) -> list[Any]:
     """
     Calculate all Euclidean distances between a selected base descriptor
     and all other Haralick descriptors
@@ -271,7 +349,7 @@ def main():
     # Number of images to perform methods on
     b_number = int(input())
 
-    files, descriptors = (np.array([]), np.array([]))
+    files, descriptors = ([], [])
 
     for _ in range(b_number):
         file = input().rstrip()
@@ -293,7 +371,7 @@ def main():
 
     # Transform ordered distances array into a sequence of indexes
     # corresponding to original file position
-    distances = get_distances(descriptors, index)
+    distances = get_distances(np.array(descriptors), index)
     indexed_distances = np.array(distances).astype(np.uint8)[:, 0]
 
     # Finally, print distances considering the Haralick descriptions from the base

From eca7118ed98f92afa8b932def0712677de873f92 Mon Sep 17 00:00:00 2001
From: "pre-commit-ci[bot]"
 <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Date: Sun, 27 Nov 2022 16:49:55 +0000
Subject: [PATCH 05/17] [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci
---
 computer_vision/haralick_descriptors.py | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)

diff --git a/computer_vision/haralick_descriptors.py b/computer_vision/haralick_descriptors.py
index 8cf0d702ce32..44c23e3f05a8 100644
--- a/computer_vision/haralick_descriptors.py
+++ b/computer_vision/haralick_descriptors.py
@@ -23,7 +23,9 @@ def root_mean_square_error(original: np.ndarray, reference: np.ndarray) -> float
     return np.sqrt(((original - reference) ** 2).mean())
 
 
-def normalize_image(image: np.ndarray, cap: float = 255.0, data_type=np.uint8) -> np.ndarray:
+def normalize_image(
+    image: np.ndarray, cap: float = 255.0, data_type=np.uint8
+) -> np.ndarray:
     """
     Normalizes image in Numpy 2D array format, between ranges 0-cap,
     as to fit uint8 type.
@@ -142,8 +144,8 @@ def transform(image: np.ndarray, kind: str, kernel=None) -> np.ndarray:
     for x in range(center_x, padded.shape[0] - center_x):
         for y in range(center_y, padded.shape[1] - center_y):
             center = padded[
-                     x - center_x: x + center_x + 1, y - center_y: y + center_y + 1
-                     ]
+                x - center_x : x + center_x + 1, y - center_y : y + center_y + 1
+            ]
             # Apply transformation method to the centered section of the image
             transformed[x - center_x, y - center_y] = apply(center[kernel == 1])
 
@@ -187,7 +189,7 @@ def closing_filter(image: np.ndarray, kernel: np.ndarray = None) -> np.ndarray:
 
 
 def binary_mask(
-        image_gray: np.ndarray, image_map: np.ndarray
+    image_gray: np.ndarray, image_map: np.ndarray
 ) -> tuple[np.ndarray, np.ndarray]:
     """
     Apply binary mask, or thresholding based
@@ -214,9 +216,7 @@ def binary_mask(
     return true_mask, false_mask
 
 
-def matrix_concurrency(
-        image: np.ndarray, coordinate
-) -> np.ndarray:
+def matrix_concurrency(image: np.ndarray, coordinate) -> np.ndarray:
     """
     Calculate sample co-occurrence matrix based on input image
     as well as selected coordinates on image.
@@ -253,7 +253,7 @@ def haralick_descriptors(matrix: np.ndarray) -> list:
     """
     # Function np.indices could be used for bigger input types,
     # but np.ogrid works just fine
-    i, j = np.ogrid[0: matrix.shape[0], 0: matrix.shape[1]]  # np.indices()
+    i, j = np.ogrid[0 : matrix.shape[0], 0 : matrix.shape[1]]  # np.indices()
 
     # Pre-calculate frequent multiplication and subtraction
     prod = np.multiply(i, j)

From 9d2cca7441d9e5351ed5b915b89afc87ad9ca989 Mon Sep 17 00:00:00 2001
From: Rafael Zimmer <rzimmerdev@gmail.com>
Date: Sun, 27 Nov 2022 13:54:33 -0300
Subject: [PATCH 06/17] Fixed line size formatting

---
 computer_vision/haralick_descriptors.py | 30 ++++++++++++++-----------
 1 file changed, 17 insertions(+), 13 deletions(-)

diff --git a/computer_vision/haralick_descriptors.py b/computer_vision/haralick_descriptors.py
index 8cf0d702ce32..b85058bc7b60 100644
--- a/computer_vision/haralick_descriptors.py
+++ b/computer_vision/haralick_descriptors.py
@@ -2,7 +2,7 @@
 https://en.wikipedia.org/wiki/Image_texture
 https://en.wikipedia.org/wiki/Co-occurrence_matrix#Application_to_image_analysis
 """
-from typing import Any, Union
+from typing import Any
 
 import imageio.v2 as imageio
 import numpy as np
@@ -23,7 +23,9 @@ def root_mean_square_error(original: np.ndarray, reference: np.ndarray) -> float
     return np.sqrt(((original - reference) ** 2).mean())
 
 
-def normalize_image(image: np.ndarray, cap: float = 255.0, data_type=np.uint8) -> np.ndarray:
+def normalize_image(
+    image: np.ndarray, cap: float = 255.0, data_type=np.uint8
+) -> np.ndarray:
     """
     Normalizes image in Numpy 2D array format, between ranges 0-cap,
     as to fit uint8 type.
@@ -36,7 +38,8 @@ def normalize_image(image: np.ndarray, cap: float = 255.0, data_type=np.uint8) -
         return 2D numpy array of type uint8, corresponding to limited range matrix
 
     Examples:
-        >>> normalize_image(np.array([[1, 2, 3], [4, 5, 10]]), cap=1.0, data_type=np.float64)
+        >>> normalize_image(np.array([[1, 2, 3], [4, 5, 10]]),
+        ...                 cap=1.0, data_type=np.float64)
         array([[0.        , 0.11111111, 0.22222222],
                [0.33333333, 0.44444444, 1.        ]])
         >>> normalize_image(np.array([[4, 4, 3], [1, 7, 2]]))
@@ -74,7 +77,8 @@ def grayscale(image: np.ndarray) -> np.ndarray:
     taking the dot product between the channel and the weights.
 
     Example:
-        >>> grayscale(np.array([[[108, 201, 72], [255, 11,  127]], [[56,  56,  56], [128, 255, 107]]]))
+        >>> grayscale(np.array([[[108, 201, 72], [255, 11,  127]],
+        ...                     [[56,  56,  56], [128, 255, 107]]]))
         array([[158,  97],
                [ 56, 200]], dtype=uint8)
     """
@@ -110,7 +114,8 @@ def transform(image: np.ndarray, kind: str, kernel=None) -> np.ndarray:
               to be used when applying convolution to original image
 
     Returns:
-        returns a numpy array with same shape as input image, corresponding to applied binary transformation.
+        returns a numpy array with same shape as input image,
+        corresponding to applied binary transformation.
 
     Examples:
         >>> img = np.array([[1, 0.5], [0.2, 0.7]])
@@ -142,8 +147,8 @@ def transform(image: np.ndarray, kind: str, kernel=None) -> np.ndarray:
     for x in range(center_x, padded.shape[0] - center_x):
         for y in range(center_y, padded.shape[1] - center_y):
             center = padded[
-                     x - center_x: x + center_x + 1, y - center_y: y + center_y + 1
-                     ]
+                x - center_x : x + center_x + 1, y - center_y : y + center_y + 1
+            ]
             # Apply transformation method to the centered section of the image
             transformed[x - center_x, y - center_y] = apply(center[kernel == 1])
 
@@ -187,7 +192,7 @@ def closing_filter(image: np.ndarray, kernel: np.ndarray = None) -> np.ndarray:
 
 
 def binary_mask(
-        image_gray: np.ndarray, image_map: np.ndarray
+    image_gray: np.ndarray, image_map: np.ndarray
 ) -> tuple[np.ndarray, np.ndarray]:
     """
     Apply binary mask, or thresholding based
@@ -196,7 +201,8 @@ def binary_mask(
     Returns the mapped true value mask and its complementary false value mask.
 
     Example:
-        >>> img = np.array([[[108, 201, 72], [255, 11,  127]], [[56,  56,  56], [128, 255, 107]]])
+        >>> img = np.array([[[108, 201, 72], [255, 11,  127]],
+        ...                 [[56,  56,  56], [128, 255, 107]]])
         >>> gray = grayscale(img)
         >>> binary = binarize(gray)
         >>> morphological = opening_filter(binary)
@@ -214,9 +220,7 @@ def binary_mask(
     return true_mask, false_mask
 
 
-def matrix_concurrency(
-        image: np.ndarray, coordinate
-) -> np.ndarray:
+def matrix_concurrency(image: np.ndarray, coordinate) -> np.ndarray:
     """
     Calculate sample co-occurrence matrix based on input image
     as well as selected coordinates on image.
@@ -253,7 +257,7 @@ def haralick_descriptors(matrix: np.ndarray) -> list:
     """
     # Function np.indices could be used for bigger input types,
     # but np.ogrid works just fine
-    i, j = np.ogrid[0: matrix.shape[0], 0: matrix.shape[1]]  # np.indices()
+    i, j = np.ogrid[0 : matrix.shape[0], 0 : matrix.shape[1]]  # np.indices()
 
     # Pre-calculate frequent multiplication and subtraction
     prod = np.multiply(i, j)

From 5e73aa6cc20e01827a0fee673c0df06770502583 Mon Sep 17 00:00:00 2001
From: Rafael Zimmer <rzimmerdev@gmail.com>
Date: Sun, 27 Nov 2022 14:21:10 -0300
Subject: [PATCH 07/17] Added final doctests

---
 computer_vision/haralick_descriptors.py | 61 +++++++++++++++++++++----
 1 file changed, 53 insertions(+), 8 deletions(-)

diff --git a/computer_vision/haralick_descriptors.py b/computer_vision/haralick_descriptors.py
index b85058bc7b60..19272673412c 100644
--- a/computer_vision/haralick_descriptors.py
+++ b/computer_vision/haralick_descriptors.py
@@ -24,7 +24,7 @@ def root_mean_square_error(original: np.ndarray, reference: np.ndarray) -> float
 
 
 def normalize_image(
-    image: np.ndarray, cap: float = 255.0, data_type=np.uint8
+    image: np.ndarray, cap: float = 255.0, data_type: np.dtype = np.uint8
 ) -> np.ndarray:
     """
     Normalizes image in Numpy 2D array format, between ranges 0-cap,
@@ -68,7 +68,8 @@ def normalize_array(array: np.ndarray, cap: float = 1) -> np.ndarray:
                [0.75],
                [1.  ]])
     """
-    return (array - np.min(array)) / (np.max(array) - np.min(array)) * cap
+    diff = np.max(array) - np.min(array)
+    return (array - np.min(array)) / (1 if diff == 0 else diff) * cap
 
 
 def grayscale(image: np.ndarray) -> np.ndarray:
@@ -101,7 +102,7 @@ def binarize(image: np.ndarray, threshold: float = 127.0) -> np.ndarray:
     return np.where(image > threshold, 1, 0)
 
 
-def transform(image: np.ndarray, kind: str, kernel=None) -> np.ndarray:
+def transform(image: np.ndarray, kind: str, kernel: np.ndarray = None) -> np.ndarray:
     """
     Simple image transformation using one of two available filter functions:
     Erosion and Dilation.
@@ -220,7 +221,7 @@ def binary_mask(
     return true_mask, false_mask
 
 
-def matrix_concurrency(image: np.ndarray, coordinate) -> np.ndarray:
+def matrix_concurrency(image: np.ndarray, coordinate: list) -> np.ndarray:
     """
     Calculate sample co-occurrence matrix based on input image
     as well as selected coordinates on image.
@@ -228,6 +229,17 @@ def matrix_concurrency(image: np.ndarray, coordinate) -> np.ndarray:
     Implementation is made using basic iteration,
     as function to be performed (np.max) is non-linear and therefore
     not callable on the frequency domain.
+
+    Example:
+        >>> img = np.array([[[108, 201, 72], [255, 11,  127]],
+        ...                 [[56,  56,  56], [128, 255, 107]]])
+        >>> gray = grayscale(img)
+        >>> binary = binarize(gray)
+        >>> morphological = opening_filter(binary)
+        >>> mask_1 = binary_mask(gray, morphological)[0]
+        >>> matrix_concurrency(mask_1, [0, 1])
+        array([[0., 0.],
+               [0., 0.]])
     """
     matrix = np.zeros([np.max(image) + 1, np.max(image) + 1])
 
@@ -239,8 +251,8 @@ def matrix_concurrency(image: np.ndarray, coordinate) -> np.ndarray:
             offset_pixel = image[x + offset_x, y + offset_y]
 
             matrix[base_pixel, offset_pixel] += 1
-
-    return matrix / np.sum(matrix)
+    matrix_sum = np.sum(matrix)
+    return matrix / (1 if matrix_sum == 0 else matrix_sum)
 
 
 def haralick_descriptors(matrix: np.ndarray) -> list:
@@ -254,6 +266,17 @@ def haralick_descriptors(matrix: np.ndarray) -> list:
 
     Returns:
         Reverse ordered list of resulting descriptors
+
+    Example:
+        >>> img = np.array([[[108, 201, 72], [255, 11,  127]],
+        ...                 [[56,  56,  56], [128, 255, 107]]])
+        >>> gray = grayscale(img)
+        >>> binary = binarize(gray)
+        >>> morphological = opening_filter(binary)
+        >>> mask_1 = binary_mask(gray, morphological)[0]
+        >>> concurrency = matrix_concurrency(mask_1, [0, 1])
+        >>> haralick_descriptors(concurrency)
+        [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]
     """
     # Function np.indices could be used for bigger input types,
     # but np.ogrid works just fine
@@ -289,10 +312,21 @@ def haralick_descriptors(matrix: np.ndarray) -> list:
     ]
 
 
-def get_descriptors(masks: tuple[np.ndarray, np.ndarray], coordinate) -> np.ndarray:
+def get_descriptors(
+    masks: tuple[np.ndarray, np.ndarray], coordinate: list
+) -> np.ndarray:
     """
     Calculate all Haralick descriptors for a sequence of
     different co-occurrence matrices, given input masks and coordinates.
+
+    Example:
+        >>> img = np.array([[[108, 201, 72], [255, 11,  127]],
+        ...                 [[56,  56,  56], [128, 255, 107]]])
+        >>> gray = grayscale(img)
+        >>> binary = binarize(gray)
+        >>> morphological = opening_filter(binary)
+        >>> get_descriptors(binary_mask(gray, morphological), [0, 1])
+        array([0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.])
     """
     descriptors = np.zeros((len(masks), 8))
     for idx, mask in enumerate(masks):
@@ -317,7 +351,7 @@ def euclidean(point_1: np.ndarray, point_2: np.ndarray) -> np.float32:
     return np.sqrt(np.sum(np.square(point_1 - point_2)))
 
 
-def get_distances(descriptors, base) -> list[Any]:
+def get_distances(descriptors: np.ndarray, base: int) -> list[Any]:
     """
     Calculate all Euclidean distances between a selected base descriptor
     and all other Haralick descriptors
@@ -331,6 +365,17 @@ def get_distances(descriptors, base) -> list[Any]:
 
     Returns:
         Ordered distances between descriptors
+
+    Example:
+        >>> index = 1
+        >>> img = np.array([[[108, 201, 72], [255, 11,  127]],
+        ...                 [[56,  56,  56], [128, 255, 107]]])
+        >>> gray = grayscale(img)
+        >>> binary = binarize(gray)
+        >>> morphological = opening_filter(binary)
+        >>> get_distances(get_descriptors(binary_mask(gray, morphological), [0, 1]), index)
+        [(0, 0.0), (1, 0.0), (2, 0.0), (3, 0.0), (4, 0.0), (5, 0.0), (6, 0.0), (7, 0.0), (8, 0.0), \
+(9, 0.0), (10, 0.0), (11, 0.0), (12, 0.0), (13, 0.0), (14, 0.0), (15, 0.0)]
     """
     distances = np.zeros(descriptors.shape[0])
 

From aacb27e201d54445f073c41354456edbc042989a Mon Sep 17 00:00:00 2001
From: Rafael Zimmer <rzimmerdev@gmail.com>
Date: Sun, 27 Nov 2022 14:28:33 -0300
Subject: [PATCH 08/17] Changed main callable

---
 computer_vision/haralick_descriptors.py | 15 +++++++--------
 1 file changed, 7 insertions(+), 8 deletions(-)

diff --git a/computer_vision/haralick_descriptors.py b/computer_vision/haralick_descriptors.py
index 19272673412c..945de2d97023 100644
--- a/computer_vision/haralick_descriptors.py
+++ b/computer_vision/haralick_descriptors.py
@@ -373,9 +373,12 @@ def get_distances(descriptors: np.ndarray, base: int) -> list[Any]:
         >>> gray = grayscale(img)
         >>> binary = binarize(gray)
         >>> morphological = opening_filter(binary)
-        >>> get_distances(get_descriptors(binary_mask(gray, morphological), [0, 1]), index)
-        [(0, 0.0), (1, 0.0), (2, 0.0), (3, 0.0), (4, 0.0), (5, 0.0), (6, 0.0), (7, 0.0), (8, 0.0), \
-(9, 0.0), (10, 0.0), (11, 0.0), (12, 0.0), (13, 0.0), (14, 0.0), (15, 0.0)]
+        >>> get_distances(get_descriptors(
+        ...                 binary_mask(gray, morphological), [0, 1]),
+        ...               index)
+        [(0, 0.0), (1, 0.0), (2, 0.0), (3, 0.0), (4, 0.0), (5, 0.0), \
+(6, 0.0), (7, 0.0), (8, 0.0), (9, 0.0), (10, 0.0), (11, 0.0), (12, 0.0), \
+(13, 0.0), (14, 0.0), (15, 0.0)]
     """
     distances = np.zeros(descriptors.shape[0])
 
@@ -386,7 +389,7 @@ def get_distances(descriptors: np.ndarray, base: int) -> list[Any]:
     return sorted(enumerate(distances), key=lambda tup: tup[1])
 
 
-def main():
+if __name__ == "__main__":
     # Index to compare haralick descriptors to
     index = int(input())
     q_value = [int(value) for value in input().split()]
@@ -429,7 +432,3 @@ def main():
     print("Ranking:")
     for idx, file_idx in enumerate(indexed_distances):
         print(f"({idx}) {files[file_idx]}", end="\n")
-
-
-if __name__ == "__main__":
-    main()

From 5fb933d057e4f28daa711dae4b947dd39d2fa551 Mon Sep 17 00:00:00 2001
From: Rafael Zimmer <rzimmerdev@gmail.com>
Date: Sun, 27 Nov 2022 14:29:21 -0300
Subject: [PATCH 09/17] Updated requirements.txt

---
 requirements.txt | 1 +
 1 file changed, 1 insertion(+)

diff --git a/requirements.txt b/requirements.txt
index a1d607df07e1..038e1cb3dbf1 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -20,3 +20,4 @@ texttable
 tweepy
 xgboost
 yulewalker
+imageio

From c9ec63ef7dce4448021a3d0a46b77a2fa0f9db59 Mon Sep 17 00:00:00 2001
From: "pre-commit-ci[bot]"
 <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Date: Sun, 27 Nov 2022 17:30:39 +0000
Subject: [PATCH 10/17] [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci
---
 requirements.txt | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/requirements.txt b/requirements.txt
index 038e1cb3dbf1..bb060edb5eba 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,5 +1,6 @@
 beautifulsoup4
 fake_useragent
+imageio
 keras
 lxml
 matplotlib
@@ -20,4 +21,3 @@ texttable
 tweepy
 xgboost
 yulewalker
-imageio

From bfd0fd7e173e562201f887ec5fe64b3b914b8f56 Mon Sep 17 00:00:00 2001
From: Rafael Zimmer <rzimmerdev@gmail.com>
Date: Tue, 13 Dec 2022 17:51:14 -0300
Subject: [PATCH 11/17] Update computer_vision/haralick_descriptors.py

No! What if the Kernel is empty?
Example:

>>> kernel = np.zeros((1))
>>> kernel or np.ones((3, 3))
array([[1., 1., 1.],
          [1., 1., 1.],
          [1., 1., 1.]])

Co-authored-by: Christian Clauss <cclauss@me.com>
---
 computer_vision/haralick_descriptors.py | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/computer_vision/haralick_descriptors.py b/computer_vision/haralick_descriptors.py
index 945de2d97023..2e1036a3e236 100644
--- a/computer_vision/haralick_descriptors.py
+++ b/computer_vision/haralick_descriptors.py
@@ -186,9 +186,7 @@ def closing_filter(image: np.ndarray, kernel: np.ndarray = None) -> np.ndarray:
         array([[0, 0],
                [0, 0]], dtype=uint8)
     """
-    if kernel is None:
-        np.ones((3, 3))
-
+    kernel = kernel or np.ones((3, 3))
     return transform(transform(image, "erosion", kernel), "dilation", kernel)
 
 

From 6f52bff54a76f7e9572d25a68e6a3860ddeb0455 Mon Sep 17 00:00:00 2001
From: Rafael Zimmer <rzimmerdev@gmail.com>
Date: Tue, 13 Dec 2022 17:52:38 -0300
Subject: [PATCH 12/17] Undone wrong commit

---
 computer_vision/haralick_descriptors.py | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/computer_vision/haralick_descriptors.py b/computer_vision/haralick_descriptors.py
index 2e1036a3e236..db9c7a7e869d 100644
--- a/computer_vision/haralick_descriptors.py
+++ b/computer_vision/haralick_descriptors.py
@@ -186,7 +186,8 @@ def closing_filter(image: np.ndarray, kernel: np.ndarray = None) -> np.ndarray:
         array([[0, 0],
                [0, 0]], dtype=uint8)
     """
-    kernel = kernel or np.ones((3, 3))
+    if kernel is None:
+        kernel = np.ones((3, 3))
     return transform(transform(image, "erosion", kernel), "dilation", kernel)
 
 

From fc2fa0bbad51e1b479c75fc96f2e8bf63a617094 Mon Sep 17 00:00:00 2001
From: Rafael Zimmer <rzimmerdev@gmail.com>
Date: Mon, 19 Dec 2022 22:43:29 -0300
Subject: [PATCH 13/17] Update haralick_descriptors.py

---
 computer_vision/haralick_descriptors.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/computer_vision/haralick_descriptors.py b/computer_vision/haralick_descriptors.py
index db9c7a7e869d..8c9afb4d0e88 100644
--- a/computer_vision/haralick_descriptors.py
+++ b/computer_vision/haralick_descriptors.py
@@ -128,7 +128,7 @@ def transform(image: np.ndarray, kind: str, kernel: np.ndarray = None) -> np.nda
         array([[0, 0],
                [0, 0]], dtype=uint8)
     """
-    if not kernel:
+    if kernel is None:
         kernel = np.ones((3, 3))
 
     if kind == "erosion":
@@ -168,7 +168,7 @@ def opening_filter(image: np.ndarray, kernel: np.ndarray = None) -> np.ndarray:
         array([[1, 1],
                [1, 1]], dtype=uint8)
     """
-    if not kernel:
+    if kernel is None:
         np.ones((3, 3))
 
     return transform(transform(image, "dilation", kernel), "erosion", kernel)

From e4c3252bb71e7a88522091a9b7efad1e12061428 Mon Sep 17 00:00:00 2001
From: Rafael Zimmer <rzimmerdev@gmail.com>
Date: Tue, 20 Dec 2022 00:33:53 -0300
Subject: [PATCH 14/17] Added algorithm for Richardson Lucy Deconvolution

---
 computer_vision/richardson_lucy.py | 233 +++++++++++++++++++++++++++++
 1 file changed, 233 insertions(+)
 create mode 100644 computer_vision/richardson_lucy.py

diff --git a/computer_vision/richardson_lucy.py b/computer_vision/richardson_lucy.py
new file mode 100644
index 000000000000..e97935735e78
--- /dev/null
+++ b/computer_vision/richardson_lucy.py
@@ -0,0 +1,233 @@
+"""
+https://en.wikipedia.org/wiki/Motion_blur
+https://en.wikipedia.org/wiki/Image_noise
+https://en.wikipedia.org/wiki/Richardson-Lucy_deconvolution
+"""
+import imageio.v2 as imageio
+import numpy as np
+from numpy.fft import fft2, ifft2, ifftshift
+
+
+def root_mean_square_error(original: np.ndarray, reference: np.ndarray) -> float:
+    """Simple implementation of Root Mean Squared Error
+    for two N dimensional numpy arrays.
+
+    Examples:
+        >>> root_mean_square_error(np.array([1, 2, 3]), np.array([1, 2, 3]))
+        0.0
+        >>> root_mean_square_error(np.array([1, 2, 3]), np.array([2, 2, 2]))
+        0.816496580927726
+        >>> root_mean_square_error(np.array([1, 2, 3]), np.array([6, 4, 2]))
+        3.1622776601683795
+    """
+    return np.sqrt(((original - reference) ** 2).mean())
+
+
+def pad_to_size(image: np.ndarray, reference: np.ndarray):
+    """Pad an image to have final shape equal to reference image."""
+    p, q = (size for size in reference.shape)
+    difference = (
+        (p - image.shape[0]) // 2,
+        (q - image.shape[1]) // 2 + (q - image.shape[1]) % 2,
+    )
+
+    return np.pad(image, difference, mode="constant")
+
+
+def normalize_image(
+    image: np.ndarray, cap: float = 255.0, data_type: np.dtype = np.uint8
+) -> np.ndarray:
+    """
+    Normalizes image in Numpy 2D array format, between ranges 0-cap,
+    as to fit uint8 type.
+
+    Args:
+        image: 2D numpy array representing image as matrix, with values in any range
+        cap: Maximum cap amount for normalization
+        data_type: numpy data type to set output variable to
+    Returns:
+        return 2D numpy array of type uint8, corresponding to limited range matrix
+
+    Examples:
+        >>> normalize_image(np.array([[1, 2, 3], [4, 5, 10]]),
+        ...                 cap=1.0, data_type=np.float64)
+        array([[0.        , 0.11111111, 0.22222222],
+               [0.33333333, 0.44444444, 1.        ]])
+        >>> normalize_image(np.array([[4, 4, 3], [1, 7, 2]]))
+        array([[127, 127,  85],
+               [  0, 255,  42]], dtype=uint8)
+    """
+    normalized = (image - np.min(image)) / (np.max(image) - np.min(image)) * cap
+    return normalized.astype(data_type)
+
+
+def gaussian_noise(size: tuple, mean=0, std=0.05):
+    """Creates normal distribution array with given size to use as noise.
+
+    Args:
+        size: Size of the desired output image
+        mean: Mean to use within the Gaussian Function
+        std: Standard deviation to use within the Gaussian Function
+
+    Returns:
+        Matrix with given size, containing generated gaussian values.
+
+    Example:
+        >>> np.random.seed(0)
+        >>> gaussian_noise((5, 5))
+        array([[ 22.49166741,   5.10200441,  12.4789093 ,  28.57138829,
+                 23.81136437],
+               [-12.46029297,  12.11362732,  -1.92980441,  -1.31604036,
+                  5.2351309 ],
+               [  1.83655553,  18.54198721,   9.703231  ,   1.55135646,
+                  5.65925622],
+               [  4.25434767,  19.04950818,  -2.61576786,   3.9916132 ,
+                -10.88972068],
+               [-32.55062015,   8.33363709,  11.02156154,  -9.46260401,
+                 28.93937146]])
+    """
+    noise = np.multiply(np.random.normal(mean, std, size), 255)
+    return noise
+
+
+def gaussian_filter(k: int = 5, sigma: float = 1.0) -> np.ndarray:
+    """Generates a matrix with weights corresponding to centered gaussian distribution.
+
+    Args:
+        k: Lateral size of the kernel.
+        sigma: Standard deviation to be used when generating distribution
+
+    Returns:
+        np.ndarray: [k x k] sized kernel to be used as filter
+
+    Examples:
+        >>> gaussian_filter()
+        array([[0.00296902, 0.01330621, 0.02193823, 0.01330621, 0.00296902],
+               [0.01330621, 0.0596343 , 0.09832033, 0.0596343 , 0.01330621],
+               [0.02193823, 0.09832033, 0.16210282, 0.09832033, 0.02193823],
+               [0.01330621, 0.0596343 , 0.09832033, 0.0596343 , 0.01330621],
+               [0.00296902, 0.01330621, 0.02193823, 0.01330621, 0.00296902]])
+    """
+    arx = np.arange((-k // 2) + 1.0, (k // 2) + 1.0)
+    x, y = np.meshgrid(arx, arx)
+    filt = np.exp(-(1 / 2) * (np.square(x) + np.square(y)) / np.square(sigma))
+    return filt / np.sum(filt)
+
+
+def convolve(matrix: np.ndarray, kernel: np.ndarray) -> np.ndarray:
+    """Convolves a given kernel around a matrix through the frequency domain, using Fourier transformations.
+
+    Args:
+        matrix: Numpy array containing values to be convolved
+        kernel: Kernel (with all dimensions smaller than those of the matrix) with weights to apply to each pixel.
+
+    Returns:
+        np.ndarray: Final equally shaped matrix with convoluted pixels.
+
+    Examples:
+        >>> matrix = np.array([[-1.45436567,  0.04575852],
+        ...                    [-0.18718385,  1.53277921]])
+        >>> kernel = np.array([[1, 0], [0, 1]])
+        >>> convolve(matrix, kernel)
+        array([[ 0.07841354, -0.14142533],
+               [-0.14142533,  0.07841354]])
+    """
+    if kernel.shape[0] > matrix.shape[1] or kernel.shape[1] > matrix.shape[1]:
+        return matrix
+    kernel = pad_to_size(kernel, matrix)
+
+    kernel_f = fft2(kernel)
+    matrix_f = fft2(matrix)
+
+    return ifftshift(ifft2(np.multiply(matrix_f, kernel_f))).real
+
+
+def get_motion_psf(shape: tuple, angle: float, num_pixel_dist: int = 20) -> np.ndarray:
+    """Generate an array with given shape corresponding to Point Spread Function for the desired angle.
+
+    Args:
+        shape: The shape of the image.
+        angle: The angle of the motion blur. Should be in degrees. [0, 360)
+        num_pixel_dist: The distance of the motion blur. [0, infinity)
+            Remember that the distance is measured in pixels. Greater will be more blurry
+
+    Returns:
+        np.ndarray: The point-spread array associated with the motion blur.
+
+    Examples:
+        >>> shape = (3, 3)
+        >>> angle = 15
+        >>> get_motion_psf(shape, angle, num_pixel_dist=3)
+        array([[0.        , 0.33333333, 0.        ],
+               [0.        , 0.33333333, 0.        ],
+               [0.33333333, 0.        , 0.        ]])
+    """
+    psf = np.zeros(shape)
+    center = np.array([shape[0] - 1, shape[1] - 1]) // 2
+    radians = angle / 180 * np.pi
+    phase = np.array([np.cos(radians), np.sin(radians)])
+    for i in range(num_pixel_dist):
+        offset_x = int(center[0] - np.round_(i * phase[0]))
+        offset_y = int(center[1] - np.round_(i * phase[1]))
+        psf[offset_x, offset_y] = 1
+    psf /= psf.sum()
+
+    return psf
+
+
+def richardson_lucy(
+    degraded: np.ndarray, function_kernel: np.ndarray, steps: int
+) -> np.ndarray:
+    """Richardson-Lucy method to restore an image affected by known motion blur function, as well as arbitrary steps
+    to perform during iterative image restoration.
+
+    Args:
+        degraded: Observed image, considered to be degraded - to be restored
+        function_kernel: Supposed function used to blur original image
+        steps: Iterative steps to take for restoring image
+
+    Returns:
+        np.ndarray: Restored image after method application
+
+
+    Examples:
+        >>> np.random.seed(0)
+        >>> shape = (3, 3)
+        >>> degraded = gaussian_noise(shape)
+        >>> kernel = np.identity(2)
+        >>> richardson_lucy(degraded, kernel, steps=4)
+        array([[1.16272775e+02, 0.00000000e+00, 7.58091075e+01],
+               [2.40582123e+01, 2.09006034e+01, 0.00000000e+00],
+               [0.00000000e+00, 1.52620152e-02, 1.31734948e+00]])
+    """
+    estimated_img = np.full(shape=degraded.shape, fill_value=1, dtype="float64")
+
+    for i in range(steps):
+        dividend = convolve(estimated_img, function_kernel)
+        quotient = np.divide(degraded, dividend, where=dividend != 0)
+
+        estimated_img = np.multiply(
+            estimated_img, convolve(quotient, np.flip(function_kernel))
+        )
+
+        estimated_img = np.clip(estimated_img, 0, 255)
+
+    return estimated_img
+
+
+def main():
+    input_file = str(input()).rstrip()
+    degraded= imageio.imread(input_file)
+
+    angle = int(input())
+    steps = int(input())
+
+    # The Richardson-Lucy iterative method is used to restore an image
+    # degraded by motion blur and noise.
+    restored = richardson_lucy(degraded, get_motion_psf(degraded.shape, angle), steps)
+
+    print(root_mean_square_error(degraded, restored))
+
+
+if __name__ == "__main__":
+    main()

From 6615ca4481d27138234568052bcf5e2bcdd1a032 Mon Sep 17 00:00:00 2001
From: Rafael Zimmer <rzimmerdev@gmail.com>
Date: Tue, 20 Dec 2022 00:39:05 -0300
Subject: [PATCH 15/17] Fixed line sizes

---
 computer_vision/richardson_lucy.py | 20 ++++++++++++++------
 1 file changed, 14 insertions(+), 6 deletions(-)

diff --git a/computer_vision/richardson_lucy.py b/computer_vision/richardson_lucy.py
index e97935735e78..1b89a1d0ee6b 100644
--- a/computer_vision/richardson_lucy.py
+++ b/computer_vision/richardson_lucy.py
@@ -115,11 +115,14 @@ def gaussian_filter(k: int = 5, sigma: float = 1.0) -> np.ndarray:
 
 
 def convolve(matrix: np.ndarray, kernel: np.ndarray) -> np.ndarray:
-    """Convolves a given kernel around a matrix through the frequency domain, using Fourier transformations.
+    """
+    Convolves a given kernel around a matrix through the frequency domain,
+    using Fourier transformations.
 
     Args:
         matrix: Numpy array containing values to be convolved
-        kernel: Kernel (with all dimensions smaller than those of the matrix) with weights to apply to each pixel.
+        kernel: Kernel (with all dimensions smaller than those of the matrix)
+                with weights to apply to each pixel.
 
     Returns:
         np.ndarray: Final equally shaped matrix with convoluted pixels.
@@ -143,13 +146,16 @@ def convolve(matrix: np.ndarray, kernel: np.ndarray) -> np.ndarray:
 
 
 def get_motion_psf(shape: tuple, angle: float, num_pixel_dist: int = 20) -> np.ndarray:
-    """Generate an array with given shape corresponding to Point Spread Function for the desired angle.
+    """
+    Generate an array with given shape corresponding to
+    Point Spread Function for the desired angle.
 
     Args:
         shape: The shape of the image.
         angle: The angle of the motion blur. Should be in degrees. [0, 360)
         num_pixel_dist: The distance of the motion blur. [0, infinity)
-            Remember that the distance is measured in pixels. Greater will be more blurry
+                        Remember that the distance is measured in pixels.
+                        Greater values will be more blurry
 
     Returns:
         np.ndarray: The point-spread array associated with the motion blur.
@@ -178,7 +184,9 @@ def get_motion_psf(shape: tuple, angle: float, num_pixel_dist: int = 20) -> np.n
 def richardson_lucy(
     degraded: np.ndarray, function_kernel: np.ndarray, steps: int
 ) -> np.ndarray:
-    """Richardson-Lucy method to restore an image affected by known motion blur function, as well as arbitrary steps
+    """
+    Richardson-Lucy method to restore an image
+    affected by known motion blur function, as well as arbitrary steps
     to perform during iterative image restoration.
 
     Args:
@@ -217,7 +225,7 @@ def richardson_lucy(
 
 def main():
     input_file = str(input()).rstrip()
-    degraded= imageio.imread(input_file)
+    degraded = imageio.imread(input_file)
 
     angle = int(input())
     steps = int(input())

From 183c8b5332339b7534dd70d99437e1218201f98e Mon Sep 17 00:00:00 2001
From: github-actions <${GITHUB_ACTOR}@users.noreply.github.com>
Date: Wed, 27 Sep 2023 06:39:02 +0000
Subject: [PATCH 16/17] updating DIRECTORY.md

---
 DIRECTORY.md | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/DIRECTORY.md b/DIRECTORY.md
index d81e4ec1ee83..289aa8bb61ca 100644
--- a/DIRECTORY.md
+++ b/DIRECTORY.md
@@ -139,6 +139,7 @@
   * [Mean Threshold](computer_vision/mean_threshold.py)
   * [Mosaic Augmentation](computer_vision/mosaic_augmentation.py)
   * [Pooling Functions](computer_vision/pooling_functions.py)
+  * [Richardson Lucy](computer_vision/richardson_lucy.py)
 
 ## Conversions
   * [Astronomical Length Scale Conversion](conversions/astronomical_length_scale_conversion.py)
@@ -515,6 +516,7 @@
   * [Logistic Regression](machine_learning/logistic_regression.py)
   * Lstm
     * [Lstm Prediction](machine_learning/lstm/lstm_prediction.py)
+  * [Mfcc](machine_learning/mfcc.py)
   * [Multilayer Perceptron Classifier](machine_learning/multilayer_perceptron_classifier.py)
   * [Polynomial Regression](machine_learning/polynomial_regression.py)
   * [Scoring Functions](machine_learning/scoring_functions.py)
@@ -672,6 +674,7 @@
   * [Sum Of Digits](maths/sum_of_digits.py)
   * [Sum Of Geometric Progression](maths/sum_of_geometric_progression.py)
   * [Sum Of Harmonic Series](maths/sum_of_harmonic_series.py)
+  * [Sum Of Two Positive Numbers Bitwise](maths/sum_of_two_positive_numbers_bitwise.py)
   * [Sumset](maths/sumset.py)
   * [Sylvester Sequence](maths/sylvester_sequence.py)
   * [Tanh](maths/tanh.py)
@@ -753,6 +756,7 @@
   * [Basic Orbital Capture](physics/basic_orbital_capture.py)
   * [Casimir Effect](physics/casimir_effect.py)
   * [Centripetal Force](physics/centripetal_force.py)
+  * [Coulombs Law](physics/coulombs_law.py)
   * [Grahams Law](physics/grahams_law.py)
   * [Horizontal Projectile Motion](physics/horizontal_projectile_motion.py)
   * [Hubble Parameter](physics/hubble_parameter.py)

From 1c040d369351705b6afe9e497ebe24d38b923335 Mon Sep 17 00:00:00 2001
From: Tianyi Zheng <tianyizheng02@gmail.com>
Date: Wed, 27 Sep 2023 02:44:21 -0400
Subject: [PATCH 17/17] Fix pre-commit errors

---
 computer_vision/richardson_lucy.py | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/computer_vision/richardson_lucy.py b/computer_vision/richardson_lucy.py
index 1b89a1d0ee6b..2087a4ab0241 100644
--- a/computer_vision/richardson_lucy.py
+++ b/computer_vision/richardson_lucy.py
@@ -173,8 +173,8 @@ def get_motion_psf(shape: tuple, angle: float, num_pixel_dist: int = 20) -> np.n
     radians = angle / 180 * np.pi
     phase = np.array([np.cos(radians), np.sin(radians)])
     for i in range(num_pixel_dist):
-        offset_x = int(center[0] - np.round_(i * phase[0]))
-        offset_y = int(center[1] - np.round_(i * phase[1]))
+        offset_x = int(center[0] - np.round(i * phase[0]))
+        offset_y = int(center[1] - np.round(i * phase[1]))
         psf[offset_x, offset_y] = 1
     psf /= psf.sum()
 
@@ -210,7 +210,7 @@ def richardson_lucy(
     """
     estimated_img = np.full(shape=degraded.shape, fill_value=1, dtype="float64")
 
-    for i in range(steps):
+    for _ in range(steps):
         dividend = convolve(estimated_img, function_kernel)
         quotient = np.divide(degraded, dividend, where=dividend != 0)