def closely(np_org_image, number_of_inters=3, struct_elem='rect', size=3):
"""Execute closely (erosion on dilated image) on a given image
Arguments:
np_org_image {NumPy array} -- image for erosion
Keyword Arguments:
struct_elem {str} -- stuctural element used during erosion (default: {'rect'})
rect - rectange
cross - cross
size {int} -- size of the structural element. Should be odd and greater than 2 (default: {3})
number_of_inters {int} -- number of inter images generated by this function (minimum/default: {3})
Returns:
[list of NumPy arrays] -- list containing original image, specified number of inter images and final image
"""
#arguments validation
bs.validate_number_of_inters(number_of_inters, closely.__defaults__[-1])
processed_inters = 0
images = []
images.append(np_org_image)
#grayscale
np_image_2D, isConverted = bs.ensureGrayscale(np_org_image, info=True)
if isConverted:
images.append(np_image_2D)
processed_inters += 1
#binarization
np_image_bin = bn.otsuBinarization(images[-1], only_threshold=False)
processed_inters += 1
images.append(np_image_bin)
#dilate part
np_image_dil = bn.dilate(images[-1], struct_elem, size)
#calculate number of inters to erode operation
dilate_inters = math.floor((number_of_inters - processed_inters) / 2 - 1)
if dilate_inters > 0:
images += bs.generateInterImages(images[-1], np_image_dil,
processed_inters + dilate_inters,
processed_inters)
processed_inters += dilate_inters
else:
images.append(np_image_dil)
#one from final dilate image
processed_inters += 1
#erode part
np_final = bn.erode(np_image_dil, struct_elem, size)
images += bs.generateInterImages(images[-1], np_final, number_of_inters,
processed_inters)
return images
def salt_pepper_noising(np_org_image,
propability=0.05,
saltPepperRatio=0.5,
number_of_inters=1):
"""Applying salt(white pixel)/pepper(black pixel) noise on a given image
Arguments:
np_org_image {NumPy array} -- image for median filtering
Keyword Arguments:
propability {float} -- propability of noising a single pixel (default: {0.05})
saltPepperRatio {float} -- specified salt to pepper ratio (default: {0.5})
1.0 -- only salt
0.5 -- equal propability of salt and pepper
0.0 -- only pepper
number_of_inters {int} -- number of inter images generated by this function (minimum/default: {1})
Returns:
[list of NumPy arrays] -- list containing original image, specified number of inter images and final image
"""
#arguments validation
bs.validate_number_of_inters(number_of_inters,
salt_pepper_noising.__defaults__[-1])
images = []
images.append(np_org_image)
np_final = ns.saltPepperNoising(np_org_image, propability, saltPepperRatio)
images += bs.generateInterImages(images[-1], np_final, number_of_inters)
return images
def gaussian_noising(np_org_image, std_dev=0.05, mean=0, number_of_inters=1):
"""Applying gaussian noise on a given image with specified standar deviation and mean parameters
Arguments:
np_org_image {NumPy array} -- image for median filtering
Keyword Arguments:
std_dev {float} -- standard deviation of gaussian noise (default: {0.05})
mean {int} --mean of gaussian noise (default: {0})
number_of_inters {int} -- number of inter images generated by this function (minimum/default: {1})
Returns:
[list of NumPy arrays] -- list containing original image, specified number of inter images and final image
"""
#arguments validation
bs.validate_number_of_inters(number_of_inters,
gaussian_noising.__defaults__[-1])
images = []
images.append(np_org_image)
np_image_3D, isConverted = bs.ensure3D(np_org_image)
np_final = ns.gaussianNoise(np_image_3D, std_dev, mean)
if isConverted:
np_final = np_final.reshape(np_final.shape[:2])
images += bs.generateInterImages(images[-1], np_final, number_of_inters)
return images
def gamma_correction(np_org_image, gamma, number_of_inters=1):
"""Gamma correction on given image using specified gamma coefficient
Arguments:
np_org_image {NumPy array} -- image for median filtering
gamma {[float]} -- gamma coefficient. Should be greater than 0
Keyword Arguments:
number_of_inters {int} -- number of inter images generated by this function (minimum/default: {1})
Returns:
[list of NumPy arrays] -- list containing original image, specified number of inter images and final image
"""
#arguments validation
bs.validate_number_of_inters(number_of_inters,
gamma_correction.__defaults__[-1])
images = []
images.append(np_org_image)
np_final = np.empty(np_org_image.shape, dtype=np.uint8)
if bs.isColorImage(np_org_image):
np_final[:, :, 0] = fil.gammaCorrection(np_org_image[:, :, 0], gamma)
np_final[:, :, 1] = fil.gammaCorrection(np_org_image[:, :, 1], gamma)
np_final[:, :, 2] = fil.gammaCorrection(np_org_image[:, :, 2], gamma)
else:
np_final = fil.gammaCorrection(np_org_image, gamma)
images += bs.generateInterImages(images[-1], np_final, number_of_inters)
return images
def filtering(np_org_image, np_mask, number_of_inters=1):
"""Filtering on a given image using specified mask.
Arguments:
np_org_image {NumPy array} -- image for filtering
np_mask {[NumPy array]} -- filter mask. You could pass your own or use one from the predefinied:
np_LP1-4 - low pass filter masks, where 1 means filter with the "strongest" low pass effect (size: {3x3})
np_HP1-4 - high pass filter masks, where 1 means filter with the "strongest" high pass effect (size: {3x3})
Keyword Arguments:
number_of_inters {int} -- number of inter images generated by this function (minimum/default: {1})
Returns:
[list of NumPy arrays] -- list containing original image, specified number of inter images and final image
"""
#arguments validation
bs.validate_number_of_inters(number_of_inters, filtering.__defaults__[-1])
images = []
images.append(np_org_image)
np_final = np.empty(np_org_image.shape, dtype=np.uint8)
if bs.isColorImage(np_org_image):
np_final[:, :, 0] = fil.matrixFilter(np_org_image[:, :, 0], np_mask)
np_final[:, :, 1] = fil.matrixFilter(np_org_image[:, :, 1], np_mask)
np_final[:, :, 2] = fil.matrixFilter(np_org_image[:, :, 2], np_mask)
else:
np_final = fil.matrixFilter(np_org_image, np_mask)
images += bs.generateInterImages(images[-1], np_final, number_of_inters)
return images
def dilation(np_org_image, struct_elem='rect', size=3, number_of_inters=2):
"""Execute dilation on a given image
Arguments:
np_org_image {NumPy array} -- image for dilation
Keyword Arguments:
struct_elem {str} -- stuctural element used during dilation (default: {'rect'})
rect - rectange
cross - cross
size {int} -- size of the structural element. Should be odd and greater than 2 (default: {3})
number_of_inters {int} -- number of inter images generated by this function (minimum/default: {2})
Returns:
[list of NumPy arrays] -- list containing original image, specified number of inter images and final image
"""
#arguments validation
bs.validate_number_of_inters(number_of_inters, dilation.__defaults__[-1])
#variable initialization
processed_inters = 0
images = []
images.append(np_org_image)
#grayscale
np_image_2D, isConverted = bs.ensureGrayscale(np_org_image, info=True)
if isConverted:
images.append(np_image_2D)
processed_inters += 1
#binarization
np_image_bin = bn.otsuBinarization(images[-1], only_threshold=False)
processed_inters += 1
images.append(np_image_bin)
#dilation
np_final = bn.dilate(images[-1], struct_elem, size)
images += bs.generateInterImages(images[-1], np_final, number_of_inters,
processed_inters)
return images
def median_filtering(np_org_image,
struct_elem='rect',
size=3,
number_of_inters=1):
"""Median filtering on a given image .
Arguments:
np_org_image {NumPy array} -- image for median filtering
Keyword Arguments:
struct_elem {str} -- stuctural element from which median value will be calculated (default: {'rect'})
rect - rectange
cross - cross
size {int} -- size of the structural element. Should be odd and greater than 2 (default: {3})
number_of_inters {int} -- number of inter images generated by this function (minimum/default: {1})
Returns:
[list of NumPy arrays] -- list containing original image, specified number of inter images and final image
"""
#arguments validation
bs.validate_number_of_inters(number_of_inters,
median_filtering.__defaults__[-1])
images = []
images.append(np_org_image)
np_final = np.empty(np_org_image.shape, dtype=np.uint8)
if bs.isColorImage(np_org_image):
np_final[:, :, 0] = fil.medianFilter(np_org_image[:, :, 0],
struct_elem, size)
np_final[:, :, 1] = fil.medianFilter(np_org_image[:, :, 1],
struct_elem, size)
np_final[:, :, 2] = fil.medianFilter(np_org_image[:, :, 2],
struct_elem, size)
else:
np_final = fil.medianFilter(np_org_image, struct_elem, size)
images += bs.generateInterImages(images[-1], np_final, number_of_inters)
return images
def binarization(np_org_image, threshold, number_of_inters=1):
"""Perform binarization on a given image using specified threshold value
Arguments:
np_org_image {NumPy array} -- image for binarization
threshold {[type]} -- binarization threshold from range <0,255>
Keyword Arguments:
number_of_inters {int} -- number of inter images generated by this function (minimum/default: {2})
Returns:
[list of NumPy arrays] -- list containing original image, specified number of inter images and final image
"""
#arguments validation
#bs.validate_number_of_inters(number_of_inters, binarization.__defaults__[-1])
if threshold > 255 or threshold < 0:
raise ValueError(
f"Threshold value out of <0,255> range. The value was {threshold}")
processed_inters = 0
images = []
images.append(np_org_image)
#grayscale
np_image_2D, isConverted = bs.ensureGrayscale(np_image, info=True)
if isConverted:
images.append(np_image_2D)
processed_inters += 1
#binarization
np_final = bn.thresholdBinarization(np_image_2D, threshold)
#number_of_inters = 0 indicates that function was called by other api function
images += bs.generateInterImages(images[-1], np_final, number_of_inters,
processed_inters)
return images
