def halo_two_dee(frame, roll_rows=30, roll_cols=30, thresh_param=50):
"""
do the halo thing with a 2D roll.
deceptively, this is doing something rather different than the regular halo, which operates on
the flattened array.
"""
# roll the frame a little (rows)
roll_frame = np.roll(frame, roll_rows, axis=0).astype(np.int32)
# cols
roll_frame = np.roll(roll_frame, roll_cols, axis=1).astype(np.int32)
# get the difference (we'll turn this into the halo)
difference = np.abs(frame.astype(np.int32) - roll_frame).astype(np.uint8)
# use adaptive thresholding to find a mask describing the interesting regions in the difference
# mask = threshold_mask(difference, block_size=thresh_param)
mask = np.linalg.norm(difference, axis=2) > thresh_param
# get the differences
difference = primitives.color_mask(difference, mask)
# return difference
# and the original image
frame = primitives.color_mask(frame, np.logical_not(mask))
return frame + difference
def halo_two_dee(frame, roll_rows=30, roll_cols=30, thresh_param=50):
"""
do the halo thing with a 2D roll.
deceptively, this is doing something rather different than the regular halo, which operates on
the flattened array.
"""
# roll the frame a little (rows)
roll_frame = np.roll(frame, roll_rows, axis=0).astype(np.int32)
# cols
roll_frame = np.roll(roll_frame, roll_cols, axis=1).astype(np.int32)
# get the difference (we'll turn this into the halo)
difference = np.abs(frame.astype(np.int32) - roll_frame).astype(np.uint8)
# use adaptive thresholding to find a mask describing the interesting regions in the difference
# mask = threshold_mask(difference, block_size=thresh_param)
mask = np.linalg.norm(difference, axis=2) > thresh_param
# get the differences
difference = primitives.color_mask(difference, mask)
# return difference
# and the original image
frame = primitives.color_mask(frame, np.logical_not(mask))
return frame + difference
def halo(frame, roll=30, thresh_param=50):
"""
let's try and put a crazy-ass halo around things.
roll=30 and thresh_param=50 with the straight hard threshold looks good.
"""
# roll the frame a little
roll_frame = np.roll(frame, roll).astype(np.int32)
# get the difference (we'll turn this into the halo)
difference = np.abs(frame.astype(np.int32) - roll_frame).astype(np.uint8)
# print(difference.min())
# print(difference.max())
# return difference
# smooth and enhance the difference
# difference = smooth_scale(difference, sigma=2, scale=150)
# use adaptive thresholding to find a mask describing the interesting regions in the difference
# mask = threshold_mask(difference, block_size=thresh_param)
mask = np.linalg.norm(difference, axis=2) > thresh_param
# get the differences
difference = primitives.color_mask(difference, mask)
# return difference
# and the original image
frame = primitives.color_mask(frame, np.logical_not(mask))
return frame + difference
def halo(frame, roll=30, thresh_param=50):
"""
let's try and put a crazy-ass halo around things.
roll=30 and thresh_param=50 with the straight hard threshold looks good.
"""
# roll the frame a little
roll_frame = np.roll(frame, roll).astype(np.int32)
# get the difference (we'll turn this into the halo)
difference = np.abs(frame.astype(np.int32) - roll_frame).astype(np.uint8)
# print(difference.min())
# print(difference.max())
# return difference
# smooth and enhance the difference
# difference = smooth_scale(difference, sigma=2, scale=150)
# use adaptive thresholding to find a mask describing the interesting regions in the difference
# mask = threshold_mask(difference, block_size=thresh_param)
mask = np.linalg.norm(difference, axis=2) > thresh_param
# get the differences
difference = primitives.color_mask(difference, mask)
# return difference
# and the original image
frame = primitives.color_mask(frame, np.logical_not(mask))
return frame + difference
def dark_sobel(frame, proportion=0.5):
"""
replace the dark regions of the image with triple sobel edges
"""
sobel = primitives.sobel_triple(frame)
mask = morphology.binary_dilation(primitives.dark_region_mask(frame, proportion=proportion))
# keep the bright regions
new_frame = primitives.color_mask(frame, mask)
# replace dark with sobel edges
# np.logical_not(mask, out=mask)
# new_frame += primitives.color_mask(sobel*4, mask)
new_frame += primitives.color_mask(sobel*4, np.logical_not(mask))
return new_frame
def dark_sobel(frame, proportion=0.5):
"""
replace the dark regions of the image with triple sobel edges
"""
sobel = primitives.sobel_triple(frame)
mask = morphology.binary_dilation(
primitives.dark_region_mask(frame, proportion=proportion))
# keep the bright regions
new_frame = primitives.color_mask(frame, mask)
# replace dark with sobel edges
# np.logical_not(mask, out=mask)
# new_frame += primitives.color_mask(sobel*4, mask)
new_frame += primitives.color_mask(sobel * 4, np.logical_not(mask))
return new_frame
def gray_skeleton(frame):
"""
draw a skeletonized corner tree on the image using the grayscale
"""
skeleton = primitives.build_skeleton(primitives.grayscale(frame))
return primitives.color_mask(frame, np.logical_not(skeleton))
def gray_skeleton(frame):
"""
draw a skeletonized corner tree on the image using the grayscale
"""
skeleton = primitives.build_skeleton(primitives.grayscale(frame))
return primitives.color_mask(frame, np.logical_not(skeleton))
def edge_tree(frame):
"""
color tree with black edges generated using original (sharp) information
"""
# get a mask for the edges
edge_mask = primitives.not_edges(frame)
# build a color tree
frame = trees.color_tree(frame)
# now apply the mask to the tree
frame = primitives.color_mask(frame, edge_mask)
return frame
def edge_tree(frame):
"""
color tree with black edges generated using original (sharp) information
"""
# get a mask for the edges
edge_mask = primitives.not_edges(frame)
# build a color tree
frame = trees.color_tree(frame)
# now apply the mask to the tree
frame = primitives.color_mask(frame, edge_mask)
return frame
def not_cell_shading(frame):
"""
this is definitely not cell shading.
"""
# get a mask for the edges
edge_mask = primitives.not_edges(frame)
# build a color tree
frame = trees.color_tree(frame)
# smooth the color tree
frame = filters.gaussian(frame, sigma=3, multichannel=True)
# now apply the mask to the tree
frame = primitives.color_mask(frame, edge_mask)
return frame
def not_cell_shading(frame):
"""
this is definitely not cell shading.
"""
# get a mask for the edges
edge_mask = primitives.not_edges(frame)
# build a color tree
frame = trees.color_tree(frame)
# smooth the color tree
frame = filters.gaussian(frame, sigma=3, multichannel=True)
# now apply the mask to the tree
frame = primitives.color_mask(frame, edge_mask)
return frame