Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Implemented a better BadPhotoCopy wave pattern effect. #332

Merged
merged 2 commits into from
Jul 29, 2023
Merged

Implemented a better BadPhotoCopy wave pattern effect. #332

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
6700970
Implemented a better BadPhotoCopy wave pattern effect.
kwcckw Jul 29, 2023
7d4d517
Fixed typo.
kwcckw Jul 29, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
87 changes: 36 additions & 51 deletions augraphy/augmentations/badphotocopy.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -100,34 +100,26 @@ def __repr__(self):

@staticmethod
@jit(nopython=True, cache=True)
def fill_wave_mask(img_wave, smooth_points, mask_ysize, mask_y_one_twelve_size):
def fill_wave_mask(img_wave, smooth_points):
"""Fill wavy pattern mask with value.

:param img_wave: The image of wavy pattern mask.
:type img_wave: numpy.array
:param smooth_points: Coordinates of wavy pattern.
:type smooth_points: numpy.array
:param mask_ysize: Height of mask.
:type mask_ysize: int
:param mask_y_one_twelve_size: One twelve of mask height.
:type mask_y_one_twelve_size: int
"""

# fill mask by draw wavy line across image
for (x_point, y_point) in smooth_points:
img_wave[:y_point, x_point] = 1
# additional noise, to smooth the edges between wavy mask
for y in range(int(y_point), int(mask_ysize - mask_y_one_twelve_size), 1):
if random.random() > 0.5:
img_wave[y, int(x_point)] = 1
if random.random() > 0.85:
break

def apply_wave(self, mask):
img_wave[:y_point, x_point] = 255

def apply_wave(self, mask, noise_side):
"""applies wavy pattern mask to input mask.

:param mask: The image to apply the function.
:type mask: numpy.array (numpy.uint8)
:param noise_side: The side of the image to apply the wave effect.
:type noise_side: string
"""

# rescale mask from 0 to 255
Expand Down Expand Up @@ -159,49 +151,48 @@ def apply_wave(self, mask):
smooth_points = smooth(points, 12).astype("int")
smooth_points = np.unique(smooth_points, axis=0)

self.fill_wave_mask(img_wave, smooth_points, mask_ysize, mask_y_one_twelve_size)
# fill wave mask
self.fill_wave_mask(img_wave, smooth_points)

# smoothen edge of the wave mask
blur_intensity = random.randint(2, 3)
gaussian_kernel_value = random.randint(151, 301)
# kernel must be odd
if not gaussian_kernel_value % 2:
gaussian_kernel_value += 1
ysize_small = max(100, int(mask_ysize / blur_intensity))
xsize_small = max(100, int(mask_xsize / blur_intensity))
img_wave = cv2.resize(img_wave, (xsize_small, ysize_small), interpolation=cv2.INTER_LINEAR)
img_wave = cv2.GaussianBlur(img_wave, (gaussian_kernel_value, gaussian_kernel_value), 0)
img_wave = cv2.resize(img_wave, (mask_xsize, mask_ysize), interpolation=cv2.INTER_LINEAR)

# top (noise concentrated at top edge)
if self.noise_side == "top":
mask = img_wave * mask
if noise_side == "top":
mask = 255 - cv2.multiply(img_wave, 255 - mask, scale=1 / 255)
# right (noise concentrated at right edge)
elif self.noise_side == "right":
elif noise_side == "right":
img_wave = np.rot90(img_wave, 3)
img_wave = cv2.resize(
img_wave,
(mask_xsize, mask_ysize),
interpolation=cv2.INTER_AREA,
)
mask = img_wave * mask
mask = 255 - cv2.multiply(img_wave, 255 - mask, scale=1 / 255)
# bottom (noise concentrated at bottom edge)
elif self.noise_side == "bottom":
elif noise_side == "bottom":
img_wave = np.flipud(img_wave)
mask = img_wave * mask
mask = 255 - cv2.multiply(img_wave, 255 - mask, scale=1 / 255)
# left (noise concentrated at left edge)
elif self.noise_side == "left":
elif noise_side == "left":
img_wave = np.rot90(img_wave, 1)
img_wave = cv2.resize(
img_wave,
(mask_xsize, mask_ysize),
interpolation=cv2.INTER_AREA,
)
mask = img_wave * mask
else:
img_wave = np.rot90(img_wave, random.randint(0, 3))
img_wave = cv2.resize(
img_wave,
(mask_xsize, mask_ysize),
interpolation=cv2.INTER_AREA,
)
mask = img_wave * mask

# reset 0 area to white (white is background)
mask[img_wave == 0] = 255

# blur the mask
mask = cv2.GaussianBlur(mask, (5, 5), cv2.BORDER_DEFAULT)
mask = 255 - cv2.multiply(img_wave, 255 - mask, scale=1 / 255)

return mask
return mask.astype("uint8")

def apply_augmentation(self, image):
"""applies augmentation to the input image.
Expand All @@ -220,6 +211,11 @@ def apply_augmentation(self, image):
# get image dimensions
ysize, xsize = image.shape[:2]

if self.noise_side == "random" or self.noise_side not in ["left", "top", "right", "bottom"]:
noise_side = random.choice(["left", "top", "right", "bottom"])
else:
noise_side = self.noise_side

# check if provided mask is numpy array
if isinstance(self.mask, np.ndarray):
mask = self.mask
Expand All @@ -231,7 +227,7 @@ def apply_augmentation(self, image):
else:
noise_generator = NoiseGenerator(
noise_type=self.noise_type,
noise_side=self.noise_side,
noise_side=noise_side,
)
mask = noise_generator.generate_noise(
noise_value=self.noise_value,
Expand All @@ -241,10 +237,6 @@ def apply_augmentation(self, image):
noise_concentration=self.noise_concentration,
)

# randomly rotate mask
if self.noise_side == "random":
mask = np.rot90(mask, random.randint(0, 3))

# new size after rotation
mask_ysize, mask_xsize = mask.shape[:2]

Expand Down Expand Up @@ -272,14 +264,7 @@ def apply_augmentation(self, image):
else:
wave_pattern = self.wave_pattern
if wave_pattern:
mask = self.apply_wave(mask)

# random flip mask vertically or horizontally
if self.noise_side == "random":
if random.choice([True, False]):
mask = cv2.flip(mask, 0)
if random.choice([True, False]):
mask = cv2.flip(mask, 1)
mask = self.apply_wave(mask, noise_side)

# add dotted noise effect to mask (unsmoothen)
if not blur_noise:
Expand Down
Loading