def test_basic(self):
img = np.random.randint(255, size=(200, 200))
pulses = lulu.decompose(img)
img_, areas, area_count = lulu.reconstruct(pulses, img.shape)
# Write assert this way so that we can see how many
# pixels mismatch as a percent of the total nr of pixels
assert_array_equal(img_, img)
assert_equal(np.sum(img_ != img) / float(np.prod(img.shape)) * 100,
0, "Percentage mismatch =")
def test_basic(self):
img = np.random.randint(255, size=(200, 200))
pulses = lulu.decompose(img)
img_, areas, area_count = lulu.reconstruct(pulses, img.shape)
# Write assert this way so that we can see how many
# pixels mismatch as a percent of the total nr of pixels
assert_array_equal(img_, img)
assert_equal(
np.sum(img_ != img) / float(np.prod(img.shape)) * 100, 0,
"Percentage mismatch =")
def lulu_filter(noisy,sf):
prev_smoothed = noisy
pulses = lulu.decompose(img=noisy,quiet=False,operator='LU')
smoothness_inc = 0.01
pulse_len = len(pulses.keys())
pulses
i=2
while 1-smoothness_inc > sf and i < pulse_len:
lulu_smoothed, areas, area_count = lulu.reconstruct(dict((k, pulses[k]) for k in pulses.keys()[0:i]), noisy.shape)
lulu_smoothed = noisy - lulu_smoothed
smoothness_inc = np.var(lulu_smoothed)/np.var(prev_smoothed)
prev_smoothed = lulu_smoothed
i = i + 1
return lulu_smoothed
bracket(t)
for t in ('amplitude_threshold', 'area_threshold',
'volume_threshold', 'circularity', 'lifetime'))
for tname in ('absolute_sum', 'amplitudes_one', 'replace', 'subtract',
'output_threshold', 'pulses_used'):
out += ('\n%s: %s' % (pptrait(tname), getattr(self, tname)))
return out
if __name__ == "__main__":
import sys
if len(sys.argv) >= 2 and '-UL' in sys.argv:
operator = 'UL'
sys.argv.remove('-UL')
else:
operator = 'LU'
image = load_image()
print("Decomposing using the %s operator." % operator)
if operator == 'LU':
print("Use the '-UL' flag to switch to UL.")
print()
pulses = lulu.decompose(image, operator=operator)
viewer = Viewer(pulses=pulses, image=image)
viewer.configure_traits()
self.plot_data.set_data('reconstruction', self.result[0])
self.new.request_redraw()
if __name__ == "__main__":
import sys
if len(sys.argv) >= 2 and '-UL' in sys.argv:
operator = 'UL'
sys.argv.remove('-UL')
else:
operator = 'LU'
image = (np.sin(np.linspace(-np.pi, np.pi, 100))*100).astype(int)
mask = np.random.random((100,)) > 0.8
noise = np.random.random((100,))*100 - 50
image += (mask * noise).astype(image.dtype)
image = image.reshape((1,-1))
print("Decomposing using the %s operator." % operator)
if operator == 'LU':
print("Use the '-UL' flag to switch to UL.")
print()
pulses = lulu.decompose(image, operator=operator)
viewer = Viewer1D(pulses=pulses, image=image)
viewer.configure_traits()
import sys sys.path.insert(0, '..') from demo import load_image import time import lulu img = load_image() print "Decomposing a %s matrix." % str(img.shape) tic = time.time() regions = lulu.decompose(img) toc = time.time() print "Execution time: %.2fs" % (toc - tic)
import numpy as np
import matplotlib.pyplot as plt
import os
import time
import lulu
import lulu.connected_region_handler as crh
img = load_image()
print "Decomposing a %s matrix." % str(img.shape)
tic = time.time()
regions = lulu.decompose(img.copy())
toc = time.time()
print "Execution time: %.2fs" % (toc - tic)
print "-" * 78
print "Reconstructing image...",
out, areas, area_count = lulu.reconstruct(regions, img.shape)
print "done."
print "Reconstructed from %d pulses." % sum(area_count)
print "-" * 78
plt.subplot(2, 2, 1)
plt.imshow(img, interpolation='nearest', cmap=plt.cm.gray)
plt.title('Original')
plt.subplot(2, 2, 2)
beta = 1.5 # Smoothing
min_area = 500
# -----------------------------------
from demo import load_image
import numpy as np
from scipy.cluster.vq import kmeans2
import matplotlib.pyplot as plt
import lulu
import lulu.connected_region_handler as crh
img = load_image('truck_and_apcs_small.jpg')
pulses = lulu.decompose(img)
impulse_strength = np.zeros(img.shape, dtype=int)
for area in pulses:
if area > min_area:
for cr in pulses[area]:
crh.set_array(impulse_strength, cr,
np.abs(crh.get_value(cr)), 'add')
def ICM(data, N, beta):
print("Performing ICM segmentation...")
# Initialise segmentation using kmeans
print("K-means initialisation...")
clusters, labels = kmeans2(np.ravel(data).astype(float), N)
beta = 1.5 # Smoothing
min_area = 500
# -----------------------------------
from demo import load_image
import numpy as np
from scipy.cluster.vq import kmeans2
import matplotlib.pyplot as plt
import lulu
import lulu.connected_region_handler as crh
img = load_image('truck_and_apcs_small.jpg')
pulses = lulu.decompose(img)
impulse_strength = np.zeros(img.shape, dtype=int)
for area in pulses:
if area > min_area:
for cr in pulses[area]:
crh.set_array(impulse_strength, cr, np.abs(crh.get_value(cr)),
'add')
def ICM(data, N, beta):
print("Performing ICM segmentation...")
# Initialise segmentation using kmeans
print("K-means initialisation...")
clusters, labels = kmeans2(np.ravel(data).astype(float), N)
import sys
sys.path.insert(0, '..')
import numpy as np
import matplotlib.pyplot as plt
import time
import lulu
sizes = []
times = []
for n in [16, 32, 64, 128, 256, 512, 1024]:
print "DPT of size (%s, %s)" % (n, n)
x = np.random.randint(255, size=(n, n))
tic = time.time()
lulu.decompose(x)
toc = time.time()
sizes.append(n * n)
times.append(toc - tic)
plt.plot(sizes, times, '-x')
plt.title('Execution times for 2D DPT')
plt.xlabel('Size, i.e. NM for an NxM image')
plt.ylabel('Execution time')
plt.show()
crh.set_array(self.pulse_strength, cr, 1, 'add')
self.result = self.pulse_strength.copy()
tmin = int(self.pulse_strength.min())
tmax = int(self.pulse_strength.max())
self.add_trait('threshold_min',
Range(value=0, low=tmin, high=tmax, mode='slider'))
self.add_trait('threshold_max',
Range(value=tmax, low=tmin, high=tmax, mode='slider'))
@on_trait_change('threshold_min, threshold_max')
def reconstruct(self):
self.result = self.pulse_strength.copy()
mask = (self.result < self.threshold_min) | \
(self.result > self.threshold_max)
self.result[mask] = 0
self.plot_data.set_data('reconstruction', self.result)
self.new.request_redraw()
image = load_image()
pulses = lulu.decompose(image)
viewer = StrengthViewer(pulses=pulses, image=image)
viewer.configure_traits()
import sys
sys.path.insert(0, '..')
import numpy as np
import matplotlib.pyplot as plt
import time
import lulu
sizes = []
times = []
for n in [16, 32, 64, 128, 256, 512, 1024]:
print("DPT of size (%s, %s)" % (n, n))
x = np.random.randint(255, size=(n, n))
tic = time.time()
lulu.decompose(x)
toc = time.time()
sizes.append(n*n)
times.append(toc - tic)
plt.plot(sizes, times, '-x')
plt.title('Execution times for 2D DPT')
plt.xlabel('Size, i.e. NM for an NxM image')
plt.ylabel('Execution time')
plt.show()
import numpy as np
import matplotlib.pyplot as plt
import os
import time
import lulu
import lulu.connected_region_handler as crh
img = load_image()
print "Decomposing a %s matrix." % str(img.shape)
tic = time.time()
regions = lulu.decompose(img.copy())
toc = time.time()
print "Execution time: %.2fs" % (toc - tic)
print "-"*78
print "Reconstructing image...",
out, areas, area_count = lulu.reconstruct(regions, img.shape)
print "done."
print "Reconstructed from %d pulses." % sum(area_count)
print "-"*78
plt.subplot(2, 2, 1)
plt.imshow(img, interpolation='nearest', cmap=plt.cm.gray)
plt.title('Original')
plt.subplot(2, 2, 2)
for cr in self.pulses[area]:
crh.set_array(self.pulse_strength, cr, 1, 'add')
self.result = self.pulse_strength.copy()
tmin = int(self.pulse_strength.min())
tmax = int(self.pulse_strength.max())
self.add_trait('threshold_min',
Range(value=0, low=tmin, high=tmax, mode='slider'))
self.add_trait('threshold_max',
Range(value=tmax, low=tmin, high=tmax, mode='slider'))
@on_trait_change('threshold_min, threshold_max')
def reconstruct(self):
self.result = self.pulse_strength.copy()
mask = (self.result < self.threshold_min) | \
(self.result > self.threshold_max)
self.result[mask] = 0
self.plot_data.set_data('reconstruction', self.result)
self.new.request_redraw()
image = load_image()
pulses = lulu.decompose(image)
viewer = StrengthViewer(pulses=pulses, image=image)
viewer.configure_traits()
