def cell_pearsons_test(self, imagemanager, cellmanager):
"""Placeholder"""
pcc_scores = {}
for key in cellmanager.cells.keys():
cell = cellmanager.cells[key]
x0, y0, x1, y1 = cell.box
channel_1_cell = imagemanager.channel_1_image[x0:x1 + 1, y0:y1 + 1]
channel_2_cell = imagemanager.channel_2_image[x0:x1 + 1, y0:y1 + 1]
c1_overlay = mark_boundaries(img_as_int(channel_1_cell),
img_as_int(cell.cell_mask),
color=(1, 0, 1),
outline_color=None)
c2_overlay = mark_boundaries(img_as_int(channel_2_cell),
img_as_int(cell.cell_mask),
color=(1, 0, 1),
outline_color=None)
cell.cell_image = np.concatenate((c1_overlay, c2_overlay), axis=1)
cell.cell_image_overlays = np.concatenate((c1_overlay, c2_overlay),
axis=1)
pcc_scores[key] = self.pearsons_score(channel_1_cell,
channel_2_cell,
cell.cell_mask)
return pcc_scores
def phys_dist2boundary(self):
print("######################################")
print("Add Physics Param: dist_to_boundary")
print("######################################")
phys_df = pd.read_csv(self.config.OUTPUT_PATH + self.config.ROOT_NAME + '-physData.csv')
if osp.exists(self.config.OUTPUT_PATH + self.config.ROOT_NAME + '-dist2boundaryMask.tif'):
dist2boundary_masks = imread(self.config.OUTPUT_PATH + self.config.ROOT_NAME + '-dist2boundaryMask.tif')
# boundary_masks = boundary_masks // 255
else:
boundary_masks = imread(self.config.OUTPUT_PATH + self.config.ROOT_NAME + '-boundaryMask.tif')
boundary_masks = boundary_masks // 255
print("######################################")
print("Generate dist2boundary_mask")
print("######################################")
dist2boundary_masks = get_dist2boundary_mask_batch(boundary_masks)
imsave(self.config.OUTPUT_PATH + self.config.ROOT_NAME + '-dist2boundaryMask.tif',
img_as_int(dist2boundary_masks))
phys_df = add_dist_to_boundary_batch_2(phys_df, dist2boundary_masks)
phys_df.round(6).to_csv(self.config.OUTPUT_PATH + self.config.ROOT_NAME + \
'-physData.csv', index=False)
def set_image(self, params, images, background):
""" creates a strip with the cell in different images
images is a list of rgb images
background is a grayscale image to use for the masks
"""
x0, y0, x1, y1 = self.box
img = color.gray2rgb(np.zeros((x1 - x0 + 1, (len(images) + 4) * (y1 - y0 + 1))))
bx0 = 0
bx1 = x1 - x0 + 1
by0 = 0
by1 = y1 - y0 + 1
for im in images:
img[bx0:bx1, by0:by1] = im[x0 : x1 + 1, y0 : y1 + 1]
by0 = by0 + y1 - y0 + 1
by1 = by1 + y1 - y0 + 1
perim = self.perim_mask
axial = self.sept_mask
cyto = self.cyto_mask
img[bx0:bx1, by0:by1] = color.gray2rgb(background[x0 : x1 + 1, y0 : y1 + 1] * self.cell_mask)
by0 = by0 + y1 - y0 + 1
by1 = by1 + y1 - y0 + 1
img[bx0:bx1, by0:by1] = color.gray2rgb(background[x0 : x1 + 1, y0 : y1 + 1] * perim)
by0 = by0 + y1 - y0 + 1
by1 = by1 + y1 - y0 + 1
img[bx0:bx1, by0:by1] = color.gray2rgb(background[x0 : x1 + 1, y0 : y1 + 1] * cyto)
if params.find_septum:
by0 = by0 + y1 - y0 + 1
by1 = by1 + y1 - y0 + 1
img[bx0:bx1, by0:by1] = color.gray2rgb(background[x0 : x1 + 1, y0 : y1 + 1] * axial)
self.image = img_as_int(img)
def myfilter(self, data):
if self.filter == 'sobel':
return util.img_as_int(filters.sobel(data))
elif self.filter == 'otsu':
thresh = filters.threshold_otsu(data)
return util.img_as_ubyte(data > thresh)
elif self.filter == '阈值分割':
thresh = self.thresholdvalue * data.max() / 100.0
return util.img_as_ubyte(data > thresh)
elif self.filter == 'canny edge':
temp = util.img_as_ubyte(
feature.canny(data, low_threshold=30, high_threshold=40))
return temp
elif self.filter == 'watershed':
mask = util.img_as_ubyte(filters.gaussian_filter(data, 0.4))
markers = feature.canny(data, low_threshold=30, high_threshold=40)
markers = ndi.label(markers)[0]
idata = filters.rank.gradient(data, morphology.disk(1))
temp = morphology.watershed(data, markers, mask=mask)
# hsv=color.convert_colorspace(temp,'L','RGB')
# io.imshow(hsv)
return temp
elif self.filter == 'test':
data = util.img_as_ubyte(filters.median(data, morphology.disk(2)))
return data
def preprocess(img):
"""
Apply image processing functions to return a binary image
"""
# Apply thresholds
cv2.imshow('unprocessed', img)
img = filters.threshold_local(img, 3)
threshold = 0.3
idx = img > img.max() * threshold
idx2 = img < img.max() * threshold
img[idx] = 0
img[idx2] = 255
# undistorting
img = cv2.remap(img, map1, map2, interpolation=cv2.INTER_LINEAR, borderMode=cv2.BORDER_CONSTANT)
# Crop the pictures as for raw images.
img = crop(img)
cv2.imshow('proced', img)
struct = ndimage.generate_binary_structure(2, 3)
# img = ndimage.binary_dilation(img, structure=struct)
img = ndimage.binary_erosion(img, ndimage.generate_binary_structure(2, 9))
img = ndimage.binary_dilation(img, structure=struct)
# cv2.imshow('proced', util.img_as_int(img))
return util.img_as_int(img)
def generate_report(self,
image_manager,
cells_manager,
fret_manager,
path=None):
if path is None:
path = tkFileDialog.askdirectory()
path = path + os.sep + "Report Experiment"
if not os.path.exists(path + os.sep + "_fret_images"):
os.makedirs(path + os.sep + "_fret_images")
if not os.path.exists(path + os.sep + "_donor_images"):
os.makedirs(path + os.sep + "_donor_images")
if not os.path.exists(path + os.sep + "_acceptor_images"):
os.makedirs(path + os.sep + "_acceptor_images")
if not os.path.exists(path + os.sep + "_control_images"):
os.makedirs(path + os.sep + "_control_images")
if not os.path.exists(path + os.sep + "_wt_images"):
os.makedirs(path + os.sep + "_wt_images")
if not os.path.exists(path + os.sep + "_discarded_images"):
os.makedirs(path + os.sep + "_discarded_images")
self.generate_report_experiment(image_manager, cells_manager,
fret_manager, path)
imsave(path + os.sep + "heatmap.png",
img_as_int(fret_manager.fret_heatmap))
def overlay_cells(cells, image, colors):
"Overlay the edges of each individual cell in the provided image"
tmp = color.gray2rgb(image)
for k in cells.keys():
c = cells[k]
if c.selection_state == 1:
col = colors[c.color_i][:3]
for px in c.outline:
x, y = px
tmp[x, y] = col
if c.sept_mask is not None:
try:
x0, y0, x1, y1 = c.box
tmp[x0:x1, y0:y1] = mark_boundaries(tmp[x0:x1, y0:y1],
img_as_int(
c.sept_mask),
color=col)
except IndexError:
c.selection_state = -1
return tmp
def overlay_cells(cells, image, colors):
"Overlay the edges of each individual cell in the provided image"
tmp = color.gray2rgb(image)
for k in cells.keys():
c = cells[k]
if c.selection_state == 1:
col = colors[c.color_i][:3]
for px in c.outline:
x, y = px
tmp[x, y] = col
if c.sept_mask is not None:
try:
x0, y0, x1, y1 = c.box
tmp[x0:x1 + 1,
y0:y1 + 1] = mark_boundaries(tmp[x0:x1 + 1, y0:y1 + 1],
img_as_int(c.sept_mask),
color=col)
except IndexError:
c.selection_state = -1
return tmp
def preprocess_algae(img):
# Crop the pictures as for raw images.
# Apply thresholds
img = filters.threshold_adaptive(img,25)
threshold = 0.3
idx = img > img.max() * threshold
idx2 = img < img.max() * threshold
img[idx] = 255
img[idx2] = 0
return util.img_as_int(img)
def preprocess_algae_fill(img):
# Crop the pictures as for raw images.
# Apply thresholds
img = filters.threshold_adaptive(img,25)
threshold = 0.3
idx = img > img.max() * threshold
idx2 = img < img.max() * threshold
img[idx] = 255
img[idx2] = 0
img = ndimage.binary_erosion(img)
img = ndimage.binary_closing(img)
return util.img_as_int(img)
def generate_thres_mask(self):
print("######################################")
print("Generate thres mask")
print("######################################")
masks_thres = self.get_thres_mask()
masks_thres_255 = np.rint(masks_thres / \
masks_thres.max() * 255).astype(np.uint8)
imsave(
self.settings['Output path'] + self.root_name + '-thresMask.tif',
masks_thres_255)
print("######################################")
print("Generate dist2thresMask")
print("######################################")
dist_masks = img_as_int(get_dist2boundary_mask_batch(masks_thres))
imsave(self.settings['Output path'] + self.root_name + \
'-dist2thresMask.tif', dist_masks)
def mask_boundary(self):
print("######################################")
print("Generate mask_boundary")
print("######################################")
boundary_masks = self.get_boundary_mask()
# Save it using 255 and 0
boundary_masks_255 = np.rint(boundary_masks / \
boundary_masks.max() * 255).astype(np.uint8)
imsave(self.config.OUTPUT_PATH + self.config.ROOT_NAME + '-boundaryMask.tif',
boundary_masks_255)
print("######################################")
print("Generate dist2boundary_mask")
print("######################################")
dist2boundary_masks = img_as_int(get_dist2boundary_mask_batch(boundary_masks))
imsave(self.config.OUTPUT_PATH + self.config.ROOT_NAME + '-dist2boundaryMask.tif',
dist2boundary_masks)
def generate_blob_mask(self):
print("######################################")
print("Generate blob mask")
print("######################################")
masks_blob = self.get_blob_mask()
masks_blob_255 = np.rint(masks_blob / \
masks_blob.max() * 255).astype(np.uint8)
imsave(self.settings['Output path'] + self.root_name + '-blobMask.tif',
masks_blob_255)
if self.settings['Generate dist2blobMask']:
print("######################################")
print("Generate dist2blob_mask")
print("######################################")
dist2blob_masks = img_as_int(
get_dist2boundary_mask_batch(masks_blob))
imsave(self.settings['Output path'] + self.root_name + \
'-dist2blobMask.tif', dist2blob_masks)
def preprocess_foam(img):
"""
Apply image processing functions to return a binary image
"""
# Crop the pictures as for raw images.
img = crop(img)
# Apply thresholds
block_size = 5
img = filters.threshold_local(img, block_size)
threshold = 0.30
idx = img > img.max() * threshold
idx2 = img < img.max() * threshold
img[idx] = 0
img[idx2] = 255
# Dilatate to get a continous network
# of liquid films
img = ndimage.binary_dilation(img)
img = ndimage.binary_dilation(img)
return util.img_as_int(img)
def preprocess_foam(img):
"""
Apply image processing functions to return a binary image
"""
# adapt to greyscale
# img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
# Apply thresholds
img = filters.threshold_local(img, 299)
threshold = 0.80
idx = img > img.max() * threshold
idx2 = img < img.max() * threshold
img[idx] = 0
img[idx2] = 255
# Dilatate to get a continous network
# of liquid films
n_dilat = 1
for _ in range(n_dilat):
img = ndimage.binary_dilation(img)
return util.img_as_int(img)
def set_image(self, params, images):
""" creates a strip with the cell in different images
images is a list of rgb images
background is a grayscale image to use for the masks
"""
x0, y0, x1, y1 = self.box
img = gray2rgb(
np.zeros((x1 - x0 + 1, (len(images) + 4) * (y1 - y0 + 1))))
bx0 = 0
bx1 = x1 - x0 + 1
by0 = 0
by1 = y1 - y0 + 1
for im in images:
img[bx0:bx1, by0:by1] = im[x0:x1 + 1, y0:y1 + 1]
by0 = by0 + y1 - y0 + 1
by1 = by1 + y1 - y0 + 1
self.image = img_as_int(img)
def mask_53bp1_blob(self):
print("######################################")
print("Generate mask_53bp1_blob")
print("######################################")
masks_53bp1_blob = self.get_53bp1_blob_mask()
# Save it using 255 and 0
masks_53bp1_blob_255 = np.rint(masks_53bp1_blob / \
masks_53bp1_blob.max() * 255).astype(np.uint8)
imsave(
self.config.OUTPUT_PATH + self.config.ROOT_NAME +
'-53bp1BlobMask.tif', masks_53bp1_blob_255)
print("######################################")
print("Generate dist253bp1blob_mask")
print("######################################")
dist253bp1blob_masks = img_as_int(
get_dist2boundary_mask_batch(masks_53bp1_blob))
imsave(
self.config.OUTPUT_PATH + self.config.ROOT_NAME +
'-dist253bp1blobMask.tif', dist253bp1blob_masks)
def set_image(self, params, images, background):
""" creates a strip with the cell in different images
images is a list of rgb images
background is a grayscale image to use for the masks
"""
x0, y0, x1, y1 = self.box
img = color.gray2rgb(
np.zeros((x1 - x0 + 1, (len(images) + 4) * (y1 - y0 + 1))))
bx0 = 0
bx1 = x1 - x0 + 1
by0 = 0
by1 = y1 - y0 + 1
for im in images:
img[bx0:bx1, by0:by1] = im[x0:x1 + 1, y0:y1 + 1]
by0 = by0 + y1 - y0 + 1
by1 = by1 + y1 - y0 + 1
perim = self.perim_mask
axial = self.sept_mask
cyto = self.cyto_mask
img[bx0:bx1, by0:by1] = color.gray2rgb(
background[x0:x1 + 1, y0:y1 + 1] * self.cell_mask)
by0 = by0 + y1 - y0 + 1
by1 = by1 + y1 - y0 + 1
img[bx0:bx1,
by0:by1] = color.gray2rgb(background[x0:x1 + 1, y0:y1 + 1] * perim)
by0 = by0 + y1 - y0 + 1
by1 = by1 + y1 - y0 + 1
img[bx0:bx1,
by0:by1] = color.gray2rgb(background[x0:x1 + 1, y0:y1 + 1] * cyto)
if params.find_septum:
by0 = by0 + y1 - y0 + 1
by1 = by1 + y1 - y0 + 1
img[bx0:bx1, by0:by1] = color.gray2rgb(
background[x0:x1 + 1, y0:y1 + 1] * axial)
self.image = img_as_int(img)
def readInImage(path):
image = img_as_int(io.imread(path, as_grey=True))
#image = img_as_uint(io.imread(path, as_grey=True))
return image
def cell_pearsons_test_percentile(self, imagemanager, cellmanager,
percentile):
pcc_scores = {}
for key in cellmanager.cells.keys():
cell = cellmanager.cells[key]
x0, y0, x1, y1 = cell.box
channel_1_cell = imagemanager.channel_1_image[x0:x1 + 1, y0:y1 + 1]
channel_2_cell = imagemanager.channel_2_image[x0:x1 + 1, y0:y1 + 1]
c1_overlay = mark_boundaries(img_as_int(channel_1_cell),
img_as_int(cell.cell_mask),
color=(1, 0, 1),
outline_color=None)
c2_overlay = mark_boundaries(img_as_int(channel_2_cell),
img_as_int(cell.cell_mask),
color=(1, 1, 0),
outline_color=None)
cell.cell_image = np.concatenate((c1_overlay, c2_overlay), axis=1)
channel_1_pxs = channel_1_cell * cell.cell_mask
channel_1_pxs = channel_1_pxs.flatten()
channel_1_pxs = sorted(channel_1_pxs)
channel_1_pxs = np.trim_zeros(channel_1_pxs)[::-1]
channel_1_pxs = channel_1_pxs[:int(
len(channel_1_pxs) * percentile)]
channel_1_threshold = channel_1_pxs[-1]
channel_1_mask = channel_1_threshold <= (channel_1_cell *
cell.cell_mask)
channel_2_pxs = channel_2_cell * cell.cell_mask
channel_2_pxs = channel_2_pxs.flatten()
channel_2_pxs = sorted(channel_2_pxs)
channel_2_pxs = np.trim_zeros(channel_2_pxs)[::-1]
channel_2_pxs = channel_2_pxs[:int(
len(channel_2_pxs) * percentile)]
channel_2_threshold = channel_2_pxs[-1]
channel_2_mask = channel_2_threshold <= (channel_2_cell *
cell.cell_mask)
c1_overlay_1 = mark_boundaries(img_as_int(c1_overlay),
img_as_int(channel_1_mask),
color=(1, 1, 0),
outline_color=None)
c2_overlay_1 = mark_boundaries(img_as_int(c2_overlay),
img_as_int(channel_1_mask),
color=(1, 1, 0),
outline_color=None)
c1_overlay_2 = mark_boundaries(img_as_int(c1_overlay),
img_as_int(channel_2_mask),
color=(1, 1, 0),
outline_color=None)
c2_overlay_2 = mark_boundaries(img_as_int(c2_overlay),
img_as_int(channel_2_mask),
color=(1, 1, 0),
outline_color=None)
cell.cell_image_overlays = np.concatenate(
(c1_overlay_1, c2_overlay_1, c1_overlay_2, c2_overlay_2),
axis=1)
#missing creation of mask for each channel and changes to allow the
#computation of two pcc's
percentile_mask = None
pcc_scores[key] = (self.pearsons_score(channel_1_cell,
channel_2_cell,
channel_1_mask),
self.pearsons_score(channel_1_cell,
channel_2_cell,
channel_2_mask))
self.cells_pcc = pcc_scores
def haralick_features(im,
prop_names=None,
distances=[2, 4, 8],
angles=np.arange(8) * np.pi / 4,
levels=256,
symmetric=False,
normed=False):
"""Compute Haralick texture features of a grayscale image.
Parameters
----------
im : 2D np.ndarray of float or uint8.
The input image.
prop_names : list of strings, optional
Texture properties of a gray level co-occurence matrix.
By default prop_names=None, which means all properties are computed.
Available texture properties include: 'contrast', 'dissimilarity',
'homogeneity', 'ASM', 'energy', and 'correlation'.
distances : array_like, optional
List of pixel pair distance offsets, used for grey covariance matrix.
angles : array_like, optional
List of pixel pair angles in radians, used for grey covariance matrix.
levels : int, optional
The input image should contain integers in [0, levels-1],
where levels indicate the number of grey-levels counted
(typically 256 for an 8-bit image).
This argument is required for 16-bit images or higher and is typically
the maximum of the image. As the output matrix is at least
levels x levels, it might be preferable to use binning of the
input image rather than large values for levels.
symmetric : bool, optional
If True, the output matrix P[:, :, d, theta] is symmetric.
This is accomplished by ignoring the order of value pairs,
so both (i, j) and (j, i) are accumulated when (i, j)
is encountered for a given offset. The default is False.
normed : bool, optional
If True, normalize each matrix P[:, :, d, theta] by dividing by
the total number of accumulated co-occurrences for the given offset.
The elements of the resulting matrix sum to 1. The default is False.
Returns
-------
fs : 1D np.ndarray of float
The feature vector.
names : list of string
The list of feature names.
References
----------
.. [1] The GLCM Tutorial Home Page,
http://www.fp.ucalgary.ca/mhallbey/tutorial.htm
"""
if np.issubdtype(im.dtype, np.floating):
im = img_as_int(im)
available_prop_names = [
'contrast', 'dissimilarity', 'homogeneity', 'ASM', 'energy',
'correlation'
]
if prop_names is None:
prop_names = available_prop_names
else: # do not allow invalid input in prop_names
prop_names = [
prop for prop in prop_names
if prop.lower() in map(str.lower, available_prop_names)
]
glcm = greycomatrix(im,
distances=distances,
angles=angles,
levels=levels,
symmetric=symmetric,
normed=normed)
fs = []
names = []
for prop in prop_names:
texture_properties = greycoprops(glcm, prop)
for dist, theta in it.product(distances, angles):
name = 'haralick-%s-distance%d-angle%d' % (prop, dist, theta)
names.append(name)
fs.append(texture_properties[distances.index(dist),
angles.index(theta)])
return np.array(fs), names
def readInImage(path):
image = img_as_int(io.imread(path, as_grey=True))
# image = img_as_uint(io.imread(path, as_grey=True))
return image
def haralick_features(im, prop_names=None, distances=[2, 4, 8], angles=np.arange(8) * np.pi/4,
levels=256, symmetric=False, normed=False):
"""Compute Haralick texture features of a grayscale image.
Parameters
----------
im : 2D np.ndarray of float or uint8.
The input image.
prop_names : list of strings, optional
Texture properties of a gray level co-occurence matrix.
By default prop_names=None, which means all properties are computed.
Available texture properties include: 'contrast', 'dissimilarity',
'homogeneity', 'ASM', 'energy', and 'correlation'.
distances : array_like, optional
List of pixel pair distance offsets, used for grey covariance matrix.
angles : array_like, optional
List of pixel pair angles in radians, used for grey covariance matrix.
levels : int, optional
The input image should contain integers in [0, levels-1],
where levels indicate the number of grey-levels counted
(typically 256 for an 8-bit image).
This argument is required for 16-bit images or higher and is typically
the maximum of the image. As the output matrix is at least
levels x levels, it might be preferable to use binning of the
input image rather than large values for levels.
symmetric : bool, optional
If True, the output matrix P[:, :, d, theta] is symmetric.
This is accomplished by ignoring the order of value pairs,
so both (i, j) and (j, i) are accumulated when (i, j)
is encountered for a given offset. The default is False.
normed : bool, optional
If True, normalize each matrix P[:, :, d, theta] by dividing by
the total number of accumulated co-occurrences for the given offset.
The elements of the resulting matrix sum to 1. The default is False.
Returns
-------
fs : 1D np.ndarray of float
The feature vector.
names : list of string
The list of feature names.
References
----------
.. [1] The GLCM Tutorial Home Page,
http://www.fp.ucalgary.ca/mhallbey/tutorial.htm
"""
if np.issubdtype(im.dtype, np.floating):
im = img_as_int(im)
available_prop_names = ['contrast',
'dissimilarity',
'homogeneity',
'ASM',
'energy',
'correlation']
if prop_names is None:
prop_names = available_prop_names
else: # do not allow invalid input in prop_names
prop_names = [prop for prop in prop_names
if prop.lower() in map(str.lower, available_prop_names)]
glcm = greycomatrix(im, distances=distances, angles=angles,
levels=levels, symmetric=symmetric, normed=normed)
fs = []
names = []
for prop in prop_names:
texture_properties = greycoprops(glcm, prop)
for dist, theta in it.product(distances, angles):
name = 'haralick-%s-distance%d-angle%d' % (prop, dist, theta)
names.append(name)
fs.append(texture_properties[distances.index(dist),
angles.index(theta)])
return np.array(fs), names
