import cv2
import numpy as np
[docs]
def normalized_variance_measure(
image: np.ndarray,
) -> float:
"""
Compute the normalized variance measure of an image.
Parameters:
image (np.ndarray): The input image as a numpy array.
Returns:
float: The computed normalized variance value.
"""
mean = np.mean(image) + 1e-8 # to avoid division by zero
var = np.var(image)
return np.divide(var, mean)
# This code is from : https://github.com/vismantic-ohtuprojekti/qualipy/blob/master/qualipy/utils/focus_measure.py
# Python implementations for focus measure operators described
# in "Analysis of focus measure operators for shape-from-focus"
# (Pattern recognition, 2012) by Pertuz et al.
# LICENSE
# The MIT License (MIT)
# Copyright (c) 2015 QualiPy developers
# Permission is hereby granted, free of charge, to any person obtaining a copy
# of this software and associated documentation files (the "Software"), to deal
# in the Software without restriction, including without limitation the rights
# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
# copies of the Software, and to permit persons to whom the Software is
# furnished to do so, subject to the following conditions:
# The above copyright notice and this permission notice shall be included in
# all copies or substantial portions of the Software.
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
# THE SOFTWARE.
[docs]
def LAPV(img: np.ndarray) -> float:
"""
Compute the Variance of Laplacian (LAP4) focus measure.
Measures the variance of the Laplacian of the image, which reflects the
amount of edges present. Higher values indicate a sharper image.
Parameters:
img (np.ndarray): The input grayscale image.
Returns:
float: Variance of the Laplacian; higher values indicate better focus.
"""
return np.std(cv2.Laplacian(img, cv2.CV_64F)) ** 2 # type: ignore
[docs]
def LAPM(img: np.ndarray) -> float:
"""
Compute the Modified Laplacian (LAP2) focus measure.
Applies separate 1D Laplacian kernels along the X and Y axes and returns
the mean of their absolute sum. Higher values indicate a sharper image.
Parameters:
img (np.ndarray): The input grayscale image.
Returns:
float: Mean Modified Laplacian; higher values indicate better focus.
"""
kernel = np.array([-1, 2, -1])
laplacianX = np.abs(cv2.filter2D(img, -1, kernel))
laplacianY = np.abs(cv2.filter2D(img, -1, kernel.T))
return np.mean(laplacianX + laplacianY)
[docs]
def TENG(img: np.ndarray) -> float:
"""
Compute the Tenengrad (TENG) focus measure.
Calculates the mean of the squared Sobel gradient magnitudes along both
axes. Higher values indicate a sharper image.
Parameters:
img (np.ndarray): The input grayscale image.
Returns:
float: Mean squared Sobel gradient; higher values indicate better focus.
"""
gaussianX = cv2.Sobel(img, cv2.CV_64F, 1, 0) # type: ignore
gaussianY = cv2.Sobel(img, cv2.CV_64F, 1, 0) # type: ignore
return np.mean(gaussianX * gaussianX + gaussianY * gaussianY)
[docs]
def MLOG(img: np.ndarray) -> float:
"""
Compute the MLOG focus measure.
Returns the maximum absolute value of the Laplacian of the image.
Higher values indicate a sharper image.
Parameters:
img (np.ndarray): The input grayscale image.
Returns:
float: Maximum absolute Laplacian; higher values indicate better focus.
"""
return np.max(cv2.convertScaleAbs(cv2.Laplacian(img, cv2.CV_64F))) # type: ignore
[docs]
def TENG_VARIANCE(img: np.ndarray) -> float:
"""
Compute the Tenengrad Variance focus measure.
Calculates the variance of the gradient magnitude (computed via Sobel
operators). Higher values indicate a sharper image.
Parameters:
img (np.ndarray): The input grayscale image.
Returns:
float: Variance of the gradient magnitude; higher values indicate better focus.
"""
gaussianX = cv2.Sobel(img, cv2.CV_64F, 1, 0) # type: ignore
gaussianY = cv2.Sobel(img, cv2.CV_64F, 1, 0) # type: ignore
G = np.sqrt(gaussianX**2 + gaussianY**2)
return np.var(G)