def main():
img = img_as_float(imread("HJoceanSmall.png"))
img_seam_v = img_as_float(imread("HJoceanSmall.png"))
img_transformed_v = img_as_float(imread("HJoceanSmall.png"))
iterations = 20
img_seam_v, img_transformed_v = seam_carve(iterations, img_seam_v, img_transformed_v)
figure()
subplot(221)
imshow(img)
title("1. Original")
subplot(222)
imshow(img_seam_v)
title("2. Seam carved vertical")
# Transposed Image
img_seam_hv = img_transformed_v.transpose(1, 0, 2)
img_transformed_hv = img_transformed_v.transpose(1, 0, 2)
iterations = 20
img_seam_hv, img_transformed_hv = seam_carve(iterations, img_seam_hv, img_transformed_hv)
subplot(223)
imshow(img_seam_hv.transpose(1, 0, 2))
title("3. Seam carved horizontal")
subplot(224)
imshow(img_transformed_hv.transpose(1, 0, 2))
title("4. Transformed Image")
show()
def AnalyseNSS(self):
if self.Mode=="Manual":
files=QFileDialog(self)
files.setWindowTitle('Non-Synchronised Segment Stripes')
self.CurrentImages=files.getOpenFileNames(self,caption='Non-Synchronised Segment Stripes')
SSSDlg1=SSSDlg.SSSWidget(self)
SSSDlg1.Img1=DCMReader.ReadDCMFile(str(self.CurrentImages[0]))
SSSDlg1.SSS1.axes.imshow(SSSDlg1.Img1,cmap='gray')
SSSDlg1.Img2=DCMReader.ReadDCMFile(str(self.CurrentImages[1]))
SSSDlg1.SSS2.axes.imshow(SSSDlg1.Img2,cmap='gray')
SSSDlg1.Img3=DCMReader.ReadDCMFile(str(self.CurrentImages[2]))
SSSDlg1.SSS3.axes.imshow(SSSDlg1.Img3,cmap='gray')
SSSDlg1.Img4=DCMReader.ReadDCMFile(str(self.CurrentImages[3]))
SSSDlg1.SSS4.axes.imshow(SSSDlg1.Img4,cmap='gray')
SSSDlg1.ImgCombi=SSSDlg1.Img1+SSSDlg1.Img2+SSSDlg1.Img3+SSSDlg1.Img4
SSSDlg1.SSSCombi.axes.imshow(SSSDlg1.ImgCombi,cmap='gray')
EPIDType=np.shape(SSSDlg1.Img1)
pl.imsave('NSS.jpg',SSSDlg1.ImgCombi)
Img1=pl.imread('NSS.jpg')
if EPIDType[0]==384:
Img2=pl.imread('NSSOrgRefas500.jpg')
else:
Img2=pl.imread('NSSOrgRef.jpg')
self.MSENSS=np.round(self.mse(Img1,Img2))
if self.Mode=="Manual":
SSSDlg1.exec_()
def find_movement():
# img = imread('shot1.jpg')
# img2 = imread('shot2.jpg')
img = imread("frame0.jpg")
img2 = imread("frame2.jpg")
img1 = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
img2 = cv2.cvtColor(img2, cv2.COLOR_BGR2GRAY)
img1 = img_as_float(img1)
img2 = img_as_float(img2)
# print img1
h1, w1 = img1.shape
h2, w2 = img2.shape
img3 = zeros((h1, w1))
for x in range(0, h1 - 1):
for y in range(0, w1 - 1):
if abs(img1[x, y] - img2[x, y]) > 0.01:
# print img1[x, y], " ", img2[x, y]
img3[x, y] = 1
figure()
# subplot(1, 2, 1), imshow(img)
# subplot(1, 2, 2), \
imshow(img3)
show()
def computeImageDifferences(numPictures):
first = rgb2gray(pl.imread("reference/0.png"))
others = []
for num in xrange(1, numPictures + 1):
others.append(rgb2gray(pl.imread("reference/%d.png" % num)))
print num
result = np.array(others) - first
#pickle.dump(result, open("computeImageDifference.pickle", "w"))
return result
def computePhaseFirstThree(folder):
import math
#transformation = np.vectorize(math.sqrt)
#transformation = np.vectorize(lambda x: x ** 2)
transformation = np.vectorize(lambda x: x) # identity
im1 = transformation(rgb2gray(pl.imread("%s/0.png" % folder)))
im2 = transformation(rgb2gray(pl.imread("%s/1.png" % folder)))
im3 = transformation(rgb2gray(pl.imread("%s/2.png" % folder)))
result = images.computePhase(im1, im2, im3)
pickle.dump(result, open("%s.pickle" % folder, "w"))
def load_img(path = 'data/rjp_small.png', gray=True):
try:
x = pylab.imread(path)
except:
x = pylab.imread('../' + path)
if len(x.shape) > 2 and gray:
x = x[:, :, 2]
if len(x.shape) > 2 and x.shape[2] == 4:
x = x[:,:,:3]
if x.max() > 1:
x = x.astype('float') / 257.0
return x
def load(self, uri):
filename = self.get(uri)
if isimg(filename):
obj = pylab.imread(filename)
elif ishdf5(filename):
f = h5py.File(filename, 'r')
obj = f[self.key(filename)].value # FIXME: lazy evaluation?
else:
try:
obj = pylab.imread(filename)
except:
raise CacheError('[bobo.cache][ERROR]: unsupported object type for loading key "%s" ' % self.key(uri))
return obj
def test_file_image(fname):
ext = os.path.splitext(fname)[-1][len(os.path.extsep):]
kwargs = to_dict_params(fname)
# Creates the image in memory
mem = BytesIO()
fractal_data = call_kw(generate_fractal, kwargs)
imsave(mem, fractal_data, cmap=kwargs["cmap"], format=ext)
mem.seek(0) # Return stream position back for reading
# Comparison pixel-by-pixel
img_file = imread("images/" + fname)
img_mem = imread(mem, format=ext)
assert img_file.tolist() == img_mem.tolist()
def load_descs(config, i, norm=1, old_desc=False, invert_mask=False, use_masks2=0):
fname = config.desc_filename(i)
mask_name = config.mask_filename(i)
mask = pylab.imread(mask_name)
descs = load_ndesc_pc(fname, norm=norm, old_desc=old_desc)
if use_masks2==0:
descs = [l for l in descs if bool(mask[l.v,l.u])!=bool(invert_mask)]
elif use_masks2==1:
mask2 = pylab.imread(config.mask2_filename(i))
descs = [l for l in descs if bool(mask2[l.v,l.u])]
elif use_masks2==2:
mask2 = pylab.imread(config.mask2_filename(i))
descs = [l for l in descs if not(mask2[l.v,l.u].astype('bool')) and mask[l.v,l.u].astype('bool')]
return descs
def myimread(imgname,flip=False,resize=None):
"""
read an image
"""
img=None
if imgname.split(".")[-1]=="png":
img=pylab.imread(imgname)
else:
img=numpy.ascontiguousarray(pylab.imread(imgname)[::-1])
if flip:
img=numpy.ascontiguousarray(img[:,::-1,:])
if resize!=None:
from scipy.misc import imresize
img=imresize(img,resize)
return img
def __init__(self, path, waveaxis=None, regions=None, start=None, end=None):
""" SpectrumImage(path[, waveaxis[, regions]]) initializes a new
spectrum image from the specified image path.
Upon initialization, the image is read in to a numpy ndarray, this
method currently assumes that 2 of the three color channels are
redundant and eliminates them. At the time of initialization, the
wavelength axis (0 - columns, 1 - rows) may be specified as well
as a tuple of lists representing regions in the form [min, max].
"""
self.image = pylab.imread(path)
self.image = self.image[:,:,0]
self.start = start
self.end = end
self.regions = []
if waveaxis is 0 or waveaxis is 1:
self.waveaxis = waveaxis
for region in regions:
bounds = self._validateregion([region['min'], region['max']])
try:
self.regions.append({'min': bounds[0], 'max': bounds[1],
'group': region['group']})
except TypeError:
pass
elif waveaxis is not None:
raise ValueError('If the wavelength axis is specified it must',
'be 0 or 1.')
def test_perspective():
# Defining the points to use. The first 4 are entered in the
# persectiveTransform function, the last is just used for the drawing of
# the yellow lines, and is the same as the first
points_x = [147, 100, 300, 392, 147]
points_y = [588, 370, 205, 392, 588]
#points_x = [570, 821, 590, 346, 570]
#points_y = [186, 170, 590, 558, 186]
# Get the image
a = imread('flyeronground.png')
a = rgb2gray(a)
subplot(121)
# Draw the lines
plot(points_x, points_y, 'y-')
imshow(a,vmin=0,vmax=1,cmap="gray")
# Calculate and show the new image
subplot(122)
b = perspectiveTransform(a, points_x[0], points_y[0],
points_x[1], points_y[1],
points_x[2], points_y[2],
points_x[3], points_y[3], 300, 200)
imshow(b,vmin=0,vmax=1,cmap="gray")
show()
def dotdraw(s, direction="RL", **kwargs):
"""
Make a drawing of an SX and display it.
direction one of "BT", "LR", "TB", "RL"
"""
try: # Check if we have pylab
from pylab import imread, imshow, show, figure, axes
except:
# We don't have pylab, so just write out to file
print "casadi.tools.graph.dotdraw: no pylab detected, will not show drawing on screen."
dotgraph(s, direction=direction, **kwargs).write_ps("temp.ps")
return
if hasattr(show, "__class__") and show.__class__.__name__ == "PylabShow":
# catch pyreport case, so we have true vector graphics
figure_name = "%s%d.%s" % (show.basename, len(show.figure_list), show.figure_extension)
show.figure_list += (figure_name,)
dotgraph(s, direction=direction, **kwargs).write_pdf(figure_name)
print "Here goes figure %s (dotdraw)" % figure_name
else:
# Matplotlib does not allow to display vector graphics on screen,
# so we fall back to png
temp = "_temp.png"
dotgraph(s, direction=direction, **kwargs).write_png(temp)
im = imread(temp)
figure()
ax = axes([0, 0, 1, 1], frameon=False)
ax.set_axis_off()
imshow(im)
show()
def get_section_image(self, cached=True):
if cached:
if self._section_image is None:
self._section_image = self.get_section_image(cached=False)
return self._section_image
else:
return pylab.imread(self.get_section_image_filename())
def main():
A = pl.imread(IMAGE_FILE)
i = 1
pc_values = (1, 5, 10, 20, 30, 40)
for num_pcs in pc_values:
# perform (truncated) pca
egvecs, proj, egvals = pca(A, num_pcs)
# reconstruct image
A_rec = np.dot(egvecs, proj).T + np.mean(A, axis=0)
# create sublplot
ax = pl.subplot(2, 3, i, frame_on=False)
ax.xaxis.set_major_locator(pl.NullLocator())
ax.yaxis.set_major_locator(pl.NullLocator())
# draw
pl.imshow(A_rec)
pl.title("{} pc's".format(num_pcs))
pl.gray()
i += 1
pl.show()
def _test():
# make a unit circle going counter-clockwise
radius = 60
theta = np.linspace(0, 2*np.pi, 10)
circle_ccw = np.array([radius*np.cos(theta), radius*np.sin(theta)])
area = ContourArea(circle_ccw)
assert area > 0
circle_cw = MakeClockwise(circle_ccw)
area = ContourArea(circle_cw)
assert area < 0
assert (circle_cw == circle_ccw[:,::-1]).all() # it actually got reversed
p = circle_cw + np.array([[280],[430]])
plb.ion()
plb.figure(0)
plb.gray()
i = plb.imread('mri2.png')
i = np.mean(i, axis=2)
plb.imshow(i)
global _contour
_contour, = plb.plot(np.append(p[0,-1], p[0]),np.append(p[1,-1], p[1]))
plb.draw()
Snake2D(i, p, iterations=500)
print 'done'
plb.ioff()
plb.savefig('mri-result.png')
plb.show()
def OnPopupItemGraph(self, event):
for row in self.ReportGrid.GetSelectedRows():
label = self.ReportGrid.GetCellValue(row,0)
id = self.ReportGrid.GetCellValue(row,1)
### plot the graph
### TODO link with properties frame
for fct in ('extTransition','intTransition', 'outputFnc', 'timeAdvance'):
filename = "%s(%s)_%s.dot"%(label,str(id),fct)
path = os.path.join(tempfile.gettempdir(), filename)
### if path exist
if os.path.exists(path):
graph = pydot.graph_from_dot_file(path)
filename_png = os.path.join(tempfile.gettempdir(),"%s(%s)_%s.png"%(label,str(id),fct))
graph.write_png(filename_png, prog='dot')
pylab.figure()
img = pylab.imread(filename_png)
pylab.imshow(img)
fig = pylab.gcf()
fig.canvas.set_window_title(filename)
pylab.axis('off')
pylab.show()
def show_fst(fst):
import pydot,pylab
graph = pydot.Dot(rankdir="LR")
isyms = fst.InputSymbols()
if not isyms: isyms = ASCII
osyms = fst.OutputSymbols()
if not osyms: osyms = ASCII
for s in range(fst.NumStates()):
if s==fst.Start():
n = pydot.Node("%d"%s,shape="box")
graph.add_node(n)
if fst.IsFinal(s):
l = '"'
l += "%d"%s # node id
if fst.Final(s).Value()!=0.0: # optional non-zero accept cost
l += "/%s"%fst.Final(s).Value()
l += '"'
n = pydot.Node("%d"%s,label=l,penwidth="3")
graph.add_node(n)
for t in range(fst.NumArcs(s)):
a = fst.GetArc(s,t)
l = '"'
l += '%s'%isyms.Find(a.ilabel)
if a.olabel!=a.ilabel: l += ":%s"%osyms.Find(a.olabel)
v = a.weight.Value()
if v!=0.0: l += "/%s"%v
l += '"'
n = a.nextstate
e = pydot.Edge("%d"%s,"%d"%n,label=l)
graph.add_edge(e)
graph.write_png("/tmp/_test.png")
pylab.gca().set_xticks([]); pylab.gca().set_yticks([])
pylab.clf()
pylab.imshow(pylab.imread("/tmp/_test.png"))
def get_neighbour_factor(path):
""" get the correct chances in the factor of an image by counting """
# load img in BW
im = np.mean(imread(path), axis=2) > 0.5
# initialize
x_diff = 0
y_diff = 0
# for each row in image starting with second
for i in range(1, im.shape[0]):
# for each individual pixel column starting with second
for j in range(1, im.shape[1]):
x_diff += abs(int(im[i-1,j]) - int(im[i,j]))
y_diff += abs(int(im[i,j-1]) - int(im[i,j]))
# divide by amount of elements with neighbours
amount_elements_x = (len(im)-1) * len(im[0])
amount_elements_y = len(im) * (len(im[0])-1)
x_diff = x_diff / float(amount_elements_x)
y_diff = y_diff / float(amount_elements_y)
print (x_diff, y_diff)
def analyze_digit_MLP(img):
""" Takes in an image matrix, crops out the digits and outputs it to file """
ocr.delete_files("../pics/cropped/")
print ("Preprocessing Image, Cropping Digits Into 28 X 28 Image Matrices\n")
cropped_img_to_show, cropped_thresh_to_Show, cropped_digits = ocr.save_digit_to_binary_img_as_mnist(img, dim = 28, saveToFile = True, imgSize = frame_new_dim)
print ("Image Preprocessing Done, %d Potential Digits Were Cropped Out\n" % len(cropped_digits))
print ("Predicting Results\n")
print ("Image Digit probability")
index = 0
for input_digit in cropped_digits:
path = "../pics/cropped/" + str(index) + ".png"
input_digit = imread(path)
digit, probability = mlp.predict(input_digit, mlp_classifier)
print ("%d.png %d %f" % (index, digit, probability))
index += 1
new_dim = (SCALE_FACTOR * img.shape[1]/2, SCALE_FACTOR * img.shape[0]/2)
cropped_img_to_show = cv2.resize(cropped_img_to_show, new_dim)
cropped_thresh_to_Show = cv2.resize(cropped_thresh_to_Show, new_dim)
cv2.imshow('handWriting Capture Cropped Image', cropped_img_to_show)
cv2.imshow('handWriting Capture Cropped Thresh', cropped_thresh_to_Show)
def main():
s = 2.0
img = imread('cameraman.png')
# Create all the images with each a differen order of convolution
img1 = gD(img, s, 0, 0)
img2 = gD(img, s, 1, 0)
img3 = gD(img, s, 0, 1)
img4 = gD(img, s, 2, 0)
img5 = gD(img, s, 0, 2)
img6 = gD(img, s, 1, 1)
fig = plt.figure()
ax1 = fig.add_subplot(2, 3, 1)
ax1.set_title("Fzero")
ax1.imshow(img1, cmap=cm.gray)
ax2 = fig.add_subplot(2, 3, 2)
ax2.set_title("Fx")
ax2.imshow(img2, cmap=cm.gray)
ax3 = fig.add_subplot(2, 3, 3)
ax3.set_title("Fy")
ax3.imshow(img3, cmap=cm.gray)
ax4 = fig.add_subplot(2, 3, 4)
ax4.set_title("Fxx")
ax4.imshow(img4, cmap=cm.gray)
ax5 = fig.add_subplot(2, 3, 5)
ax5.set_title("Fyy")
ax5.imshow(img5, cmap=cm.gray)
ax6 = fig.add_subplot(2, 3, 6)
ax6.set_title("Fxy")
ax6.imshow(img6, cmap=cm.gray)
show()
def structure_plot(fig, molids, activations=None):
"""plot molecule structures"""
if activations != None:
assert len(activations[0]) == len(molids[0])
assert len(activations[1]) == len(molids[1])
id2name = defaultdict(str, rdl.get_id2name())
all_molids = molids[0] + molids[1]
all_activations = np.hstack(activations)
cr = utils.ceiled_root(len(all_molids))
for i, molid in enumerate(all_molids):
ax = fig.add_subplot(cr, cr, i+1)
try:
img = plt.imread(os.path.join(structures_path, molid + '.png'))
except Exception, e:
img = np.zeros(img.shape)
ax.imshow(img)
ax.set_xticks([])
ax.set_yticks([])
if i >= len(molids[0]):
for child in ax.get_children():
if isinstance(child, plt.matplotlib.spines.Spine):
child.set_color('#ff0000')
if activations != None:
ax.set_title('{}: {:.2f}'.format(id2name[molid], all_activations[i]),
rotation='20')
def extract_features(image_path_list):
feature_list = []
for image_path in image_path_list:
features = []
image_array = imread(image_path)
# Note: Looping through multiple filters for edge detection drastically slows
# Down the feature extraction while only marginally improving performance, thus
# it is left out for the HW submission
# for ax in [0,1]:
# for pct in [.01, .02]:
emat = featureExtractor.getEdgeMatrix(image_array, sigpercent=.01, \
axis=0)
features.append( featureExtractor.getEdgePercent(image_array, emat) )
features.append( featureExtractor.getNumMeridialEdges(emat) )
features.append( featureExtractor.getNumEquatorialEdges(emat) )
features.append( featureExtractor.getSize(image_array) )
features.append( featureExtractor.getCentralRatio(image_array) )
features.append( featureExtractor.getCentralRatio(emat) )
features.append( featureExtractor.getMeanColorVal(image_array, 0) )
features.append( featureExtractor.getMeanColorVal(image_array, 1) )
features.append( featureExtractor.getMeanColorVal(image_array, 2) )
features.append( featureExtractor.getVariance( image_array, 0 ) )
features.append( featureExtractor.getVariance( image_array, 1 ) )
features.append( featureExtractor.getVariance( image_array, 2 ) )
xr, yr = featureExtractor.getCOM(image_array, 0)
features.append(xr)
features.append(yr)
xg, yg = featureExtractor.getCOM(image_array, 1)
features.append(xg)
features.append(yg)
xb, yb = featureExtractor.getCOM(image_array, 2)
features.append(xb)
features.append(yb)
feature_list.append([image_path, features])
return feature_list
def read_tiff(fname,res_x,res_y,pix_x,pix_y,ext_x,ext_y): # get array numbers img=pl.imread(fname) if len(img.shape)==2: img_arraynum=1 else: img_arraynum=img.shape[2] #collapse accordingly if img_arraynum == 4: imgmat=numpy.multiply(img[:,:,0]+img[:,:,1]+img[:,:,2],img[:,:,3]) elif img_arraynum == 1: imgmat = img else: print "Image has %d arrays, unhandled." % img_arraynum exit(0) ### convert data to float64 imgmat = numpy.array(imgmat,dtype=numpy.float64) ### image parameters pix_x = imgmat.shape[1] pix_y = imgmat.shape[0] ext_x = pix_x * res_x ext_y = pix_y * res_y ### convert to linear imgmat = convert2lin(imgmat,latitude,sensitivity) ### return final image return imgmat,res_x,res_y,pix_x,pix_y,ext_x,ext_y
def test_with_file(fn):
im = pylab.imread(fn)
if im.ndim > 2:
im = numpy.mean(im[:, :, :3], 2)
pylab.imsave("intermediate.png", im, vmin=0, vmax=1., cmap=pylab.cm.gray)
r = test_inline(im)
return r
def extract_features(image_path_list):
feature_list = []
for image_path in image_path_list:
features = []
image_array = imread(image_path)
#############################################################################################
# INSERT THE FEATURES THAT YOU EXTRACT FROM THE IMAGE HERE #
#############################################################################################
# for ax in [0,1]:
# for pct in [.01, .02]:
# emat = featureExtractor.getEdgeMatrix(image_array, sigpercent=pct, \
# axis=ax)
# features.append( featureExtractor.getEdgePercent(image_array, emat) )
# features.append( featureExtractor.getNumMeridialEdges(emat) )
# features.append( featureExtractor.getNumEquatorialEdges(emat) )
# features.append( featureExtractor.getSize(image_array) )
# features.append( featureExtractor.getCentralRatio(image_array) )
# features.append( featureExtractor.getCentralRatio(emat) )
# features.append( featureExtractor.getMeanColorVal(image_array, 0) )
# features.append( featureExtractor.getMeanColorVal(image_array, 1) )
# features.append( featureExtractor.getMeanColorVal(image_array, 2) )
# features.append( featureExtractor.getVariance( image_array, 0 ) )
# features.append( featureExtractor.getVariance( image_array, 1 ) )
# features.append( featureExtractor.getVariance( image_array, 2 ) )
# xr, yr = featureExtractor.getCOM(image_array, 0)
# features.append(xr)
# features.append(yr)
# xg, yg = featureExtractor.getCOM(image_array, 1)
# features.append(xg)
# features.append(yg)
# xb, yb = featureExtractor.getCOM(image_array, 2)
# features.append(xb)
# features.append(yb)
feature_list.append([image_path, features])
return feature_list
def m2screenshot(mayavi_fig=None, mpl_axes=None, autocrop=True):
""" Capture a screeshot of the Mayavi figure and display it in the
matplotlib axes.
"""
import pylab as pl
# Late import to avoid triggering wx imports before needed.
try:
from mayavi import mlab
except ImportError:
# Try out old install of Mayavi, with namespace packages
from enthought.mayavi import mlab
if mayavi_fig is None:
mayavi_fig = mlab.gcf()
else:
mlab.figure(mayavi_fig)
if mpl_axes is not None:
pl.axes(mpl_axes)
filename = tempfile.mktemp('.png')
mlab.savefig(filename, figure=mayavi_fig)
image3d = pl.imread(filename)
if autocrop:
bg_color = mayavi_fig.scene.background
image3d = autocrop_img(image3d, bg_color)
pl.imshow(image3d)
pl.axis('off')
os.unlink(filename)
def showOtherRF(self):
if self.path1 != '' and self.path2 != '':
self.calcDistance()
else:
print "Nie wybrano punktow"
# WYSWIETLA INFORMACJE
img = pylab.imread('img/wally.png', 'rb')
pylab.imshow(img)
pylab.plot(0, 0)
# DAJE CZYSTY OBRAZ BEZ OSI ** PEWNIE MOZNA PROSCIEJ
frame1 = pylab.gca()
for xlabel_i in frame1.axes.get_xticklabels():
xlabel_i.set_visible(False)
xlabel_i.set_fontsize(0.0)
for xlabel_i in frame1.axes.get_yticklabels():
xlabel_i.set_fontsize(0.0)
xlabel_i.set_visible(False)
for tick in frame1.axes.get_xticklines():
tick.set_visible(False)
for tick in frame1.axes.get_yticklines():
tick.set_visible(False)
# SHOWTIME
pylab.show()
def cropAutoWhitePNG(IMGFileName, padding = (10, 10)): assert(os.path.isfile(IMGFileName)),"PNG file does not exist" IMG = pylab.imread(IMGFileName) if IMG.shape[2] == 4: T = numpy.squeeze(numpy.take(IMG, [0, 1, 2], axis = 2)) T = numpy.any(T < 1, axis = 2) elif IMG.shape[2] == 3: T = numpy.any(T < 1, axis = 2) elif IMG.ndim < 3: T = (T < 1) I = numpy.where(T) croppedIMG = numpy.array(IMG) for z in range(2): croppedIMG = numpy.take(croppedIMG, numpy.arange(numpy.min(I[z]), numpy.max(I[z]) + 1), axis = z) cornerPadding = numpy.ones((padding[0], padding[1], croppedIMG.shape[2])) topBottomPadding = numpy.ones((padding[0], croppedIMG.shape[1], croppedIMG.shape[2])) leftRightPadding = numpy.ones((croppedIMG.shape[0], padding[1], croppedIMG.shape[2])) T = numpy.concatenate(( numpy.concatenate((cornerPadding, topBottomPadding, cornerPadding), axis = 1), numpy.concatenate((leftRightPadding, croppedIMG, leftRightPadding), axis = 1), numpy.concatenate((cornerPadding, topBottomPadding, cornerPadding), axis = 1)), axis = 0) scipy.misc.imsave(IMGFileName, T)
def show_xyfm(point_in_0, xyfm):
x0,y0 = point_in_0
for i,(x, y, f, m) in enumerate(xyfm):
pylab.figure(i);
pylab.imshow(pylab.imread(f))
xyw = np.dot(m,[x0, y0, 1])
x1,y1 = xyw[:2] / xyw[2]
pylab.scatter(x1, y1, c='r', marker='+')
import pylab as pl
from scipy.misc import imresize, imfilter
import turtle
# load image
img = pl.flipud(pl.imread(".aar.png"))
a=turtle.Turtle()
s=turtle.Screen()
s.setup(400,400)
s.delay(5000)
# setup turtle
levels = 8
size = 2**levels
turtle.setup(img.shape[1] * 4.2, img.shape[0] * 4.2)
turtle.setworldcoordinates(0, 0, size, -size)
turtle.tracer(1000, 0)
# resize and blur image
img = imfilter(imresize(img, (size, size)), 'blur')
# define recursive hilbert curve
def hilbert(level, angle = 90):
if level == 0:
return
if level == 1 and img[-turtle.pos()[1], turtle.pos()[0]] > 128:
turtle.forward(2**level - 1)
else:
turtle.right(angle)
hilbert(level - 1, -angle)
turtle.forward(1)
turtle.left(angle)
def initialize(self):
texture0 = np.zeros((100, 100, 4))
texture1 = imread(get_image_path('_REF.png'))
self.add_visual(TextureVisual, texture=texture0)
self.set_data(texture=texture1)
assert not (np.isnan(self._ll))
if self._verbose:
print self._ll
return self._ll
def mixture(self):
return self._mm
if __name__ == "__main__":
from ..families import UnivariateGaussian
from pylab import imread
import sys
k = 8
im = imread("examples/data/ar.ppm")
data = (im.sum(axis=2) / 3.0).flatten().reshape((-1, 1))
print data.shape
em = BregmanSoftClustering(data, k, UnivariateGaussian, ())
mm = em.run()
mm.savetxt(sys.stdout)
from matplotlib import pyplot
pyplot.subplot(2, 1, 1)
pyplot.hist(data, 256)
pyplot.xlim(0, 256)
pyplot.subplot(2, 1, 2)
x = np.arange(0, 255, 0.1)
pyplot.plot(x, mm(x))
patch += J[i, j] * J[i + k, j + l]
energya = -eta * np.sum(I * J) - zeta * patch
J[i, j] = -1
patch = 0
for k in range(-1, 1):
for l in range(-1, 1):
patch += J[i, j] * J[i + k, j + l]
energyb = -eta * np.sum(I * J) - zeta * patch
if energya < energyb:
J[i, j] = 1
else:
J[i, j] = -1
return J
I = pl.imread('world.png')
N = np.shape(I)[0]
I = I[:, :, 0]
I = np.where(I < 0.1, -1, 1)
pl.imshow(I)
pl.title('Original Image')
noise = np.random.rand(N, N)
J = I.copy()
ind = np.where(noise < 0.1)
J[ind] = -J[ind]
pl.figure()
pl.imshow(J)
pl.title('Noisy image')
newJ = J.copy()
newJ = MRF(I, newJ)
def make_Fig5(exc_connectivity1='random',
net_state1='asynchronous',
exc_connectivity2='random',
net_state2='synchronous',
exc_connectivity3='clustered',
net_state3='synchronous',
node_group='1_random',
N_E=800,
N_I=200,
tau_Em=50,
tau_Im=25,
p_exc=0.03,
p_inh=0.20,
tau_ref_E=10,
tau_ref_I=5,
V_leak_E_min=-70,
V_leak_I_min=-70,
V_th=-40,
V_reset=-59,
V_peak=0,
V_syn_E=50,
V_syn_I=-68,
C_E=0.4,
C_I=0.2,
g_AE1=2.5e-3,
g_AE2=2.5e-3,
g_AE3=3.0e-3,
g_AI=6.0e-3,
g_GE=30.0e-3,
g_GI=30.0e-3,
g_extE=6.0e-3,
g_extI1=4.0e-3,
g_extI2=4.0e-3,
g_extI3=6.0e-3,
g_leak_E=10.0e-3,
g_leak_I=5.0e-3,
tau_Fl=1.5,
tau_AIl=1.5,
tau_AIr=0.2,
tau_AId=1.0,
tau_AEl=1.5,
tau_AEr=0.2,
tau_AEd=1.0,
tauR_intra=2500.0,
tauD_intra=15.0,
tauR_extra=2500.0,
tauD_extra=15.0,
rateI=65,
rateE=65,
stim_factor=4.5,
num_nodes_to_save=20,
num_pairs=40,
num_LFPs=8,
windows=[1000.0, 1000.0, 1000.0],
gaps=[200.0, 0.0, 200.0],
padding_for_traces=500.0,
stim_type='linear_increase',
time_to_max=200,
inh_connectivity='random',
exc_syn_weight_dist='beta',
ext_SD=True,
int_SD=True,
generate_LFPs=True,
save_results=True,
num_trials=15,
crit_freq_V=(20.0, 100.0),
filt_kind_V='band',
crit_freq_LFP=100.0,
filt_kind_LFP='low',
V_mult_factor=10000.0,
LFP_mult_factor=-2.5,
run_new_net_sim=False,
get_new_CC_results_all_pairs=False,
master_folder_path='E:\\correlated_variability_',
showfig=True,
savefig=False):
'''
Generate Figure 5 from the manuscript. Also, calculate and report CC
results for clustered network with no synaptic depression.
Parameters
----------
exc_connectivity1: string
type of E --> E and E --> I connectivity in model network used in A
and B ('clustered' or 'random')
net_state1: string
Network 'state' for model used in A and B.
If 'synchronous', choose inh synaptic time constants that support
network spike-rate oscillations in response to the stimulus.
Otherwise, use same time constants for inh and exc synapses.
exc_connectivity2: string
exc connectivity in model network used in C and D
net_state2: string
Network 'state' for model used in C and D.
exc_connectivity3: string
exc connectivity in model network used in E and F
net_state3: string
Network 'state' for model used in E and F.
node_group: float
group of network nodes from which to select example Vs and LFP
tau_Em: float
membrane time constant in ms for exc nodes
tau_Im: float
membrane time constant in ms for inh nodes
N_E: float or int
number of exc nodes
N_I: float or int
number of inh nodes
p_exc: float
connection probability for E --> E and E --> I
p_inh: float
connection probability for I --> E and I --> I
tau_ref_E: float
absolute refractory period for exc nodes in ms
tau_ref_I: float
absolute refractory period for inh nodes in ms
V_leak_I: float or int
leak reversal potential for inh nodes in mV
V_leak_E: float or int
leak reversal potential for exc nodes in mV
V_th: float or int
spike threshold for all nodes in mV
V_reset: float or int
post-spike reset membrane potential for all nodes in mV
V_peak: float or int
peak spike membrane potential for all nodes in mV
dt: float
time step in ms
V_syn_E: float or int
synaptic reversal potential for exc nodes in mV
V_syn_I: float or int
synaptic reversal potential for inh nodes in mV
g_AE1: float
synaptic conductance for AMPA channels on exc nodes in microS
for model network used in A and B
g_AE2: float
synaptic conductance for AMPA channels on exc nodes in microS
for model network used in C and D
g_AE3: float
synaptic conductance for AMPA channels on exc nodes in microS
for model network used in E and F
g_AI: float
synaptic conductance for AMPA channels on inh nodes in microS
g_GE: float
synaptic conductance for GABA channels on exc nodes in microS
g_GI1: float
synaptic conductance for GABA channels on inh nodes in microS
for model network used in A and B
g_GI2: float
synaptic conductance for GABA channels on inh nodes in microS
for model network used in C and D
g_GI3: float
synaptic conductance for GABA channels on inh nodes in microS
for model network used in E and F
tau_Fl: float or int
inh to inh delay time constant in ms
tau_Fr: float
inh to inh rise time constant in ms
tau_Fd: float
inh to inh decay time constant in ms
tau_AIl: float
exc to inh AMPA delay time constant in ms
tau_AIr: float
exc to inh AMPA rise time constant in ms
tau_AId: float
exc to inh AMPA decay time constant in ms
tau_AEl: float
exc to exc AMPA delay time constant in ms
tau_AEr: float
exc to exc AMPA rise time constant in ms
tau_AEd: float
exc to exc AMPA decay time constant in ms
tauR_intra: float
synaptic weight recovery time constant in ms for intracortical
synapses
tauD_intra: float
synaptic weight decay time constant in ms for intracortical synapses
tauR_extra float
synaptic weight recovery time constant in ms for external inputs
tauD_extra: float
synaptic weight decay time constant in ms for external inputs
g_extE: float
synaptic conductance for ext input on exc in microS
g_extI: float
synaptic conductance for ext input on inh in microS
C_E: float
capacitance for exc nodes in nF
C_I: float
capacitance for inh nodes in nF
g_leak_E: float
leak conductance for exc nodes in microS
g_leak_I: float l
leak conductance for inh nodes in microS
rateI: float or int
ongoing external input rate for exc nodes in Hz
rateE: float or int
ongoing external input rate for inh nodes in Hz
stim_factor: float or int
factor by which external input increases after stim onset
num_nodes_to_save: int
number of nodes for which to save synaptic inputs (for injecting
into 'test' neurons in a separate function)
num_pairs: int
number of test neuron pairs for which to calculate CCs
num_LFPs: float
number of LFPs to simulate (determines # of nodes in each 'electrode')
windows: list of floats
widths of ongoing, transient, and steady-state windows (ms)
gaps: list of floats
sizes of gaps between stim onset and end of ongoing, stim onset and
beginning of transient, end of transient and beginning of
steady-state (ms)
padding: float
size of window (ms) to be added to the beginning and end of each
simulation
stim_type: string
if 'linear_increase', the stimulus is mimicked as a gradual (step-wise)
increase in external input rate. Otherwise, single step function.
time_to_max: float or int
time to reach max input rate for 'linear_increase' stimulus (in ms)
inh_connectivity: string
type of I --> E and I --> I connectivity ('clustered' or 'random')
exc_syn_weight_dist: string
type of distribution from which to draw E --> E and E --> I nonzero
weights. If 'beta', draw from beta distribution. Otherwise, draw
from continuous uniform distribution.
ext_SD: bool
if True, apply synaptic adaptation to external input synapses after
stimulus onset
int_SD: bool
if True, apply synaptic adaptation to all 'intracortical' synapses
after stimulus onset
generate_LFPs: bool
if True, simulate LFPs as sums of synaptic currents to groups of
exc nodes
save_results: bool
if True, save results for each trial
num_trials: float or int
number of trials to simulation
crit_freq_V: float or tuple of floats
critical frequency for filtering membrane potentials
(e.g., (20.0, 100.0))
filt_kind_V: string
kind of filter to use for membrane potentials (e.g., 'band' for
bandpass)
crit_freq_LFP: float or tuple of floats
critical frequency for filtering 'LFP' (e.g., 100.0)
filt_kind_LFP: string
kind of filter to use for 'LFP' (e.g., 'low' for lowpass)
V_mult_factor: int or float
factor by which to multiply membrane potentials for example plot
(e.g., 10000.0)
LFP_mult_factor: int or float
factor by which to multiply 'LFP' for example plot (e.g., -2.5)
run_new_net_sim: Bool
if True, run new network simulations (for num_trials trials), even if
results for these settings already exist
get_new_CC_results_all_pairs: Bool
if True, calculate CCs for all pairs starting with traces.
master_folder_path: string
full path of directory containing data, code, figures, etc.
showfig: Bool
if True (and if savefig == True), plot spike rasters, include
network schematics, and show the figure. Otherwise, just calculate
CCs for each of the three networks (which is relatively fast if
simulations already complete).
savefig: Bool
if True, save the figure to the specified path
Returns
-------
None
'''
width_fig = 5.0
height_fig = 9.0
padding = 0.5 #space (in inches) b/w subfigs and edges of figure
interpadding = 0.4 #reference unit for padding b/w subfigs
fonts = FontProperties()
fonts.set_weight('bold')
fonts.set_size(6)
fontm = FontProperties()
fontm.set_weight('bold')
fontm.set_size(12)
fontl = FontProperties()
fontl.set_weight('bold')
fontl.set_size(12)
mpl.rcParams['mathtext.default'] = 'regular'
#a1:spike rasters, V1-V2 (20-100 Hz), LFP (for model_version1)
#a2:stim trace for a1
#a3:model schematic
#b:V-V CC trajectories
#c1:spike rasters, V1-V2 (20-100 Hz), LFP (for model_version2)
#c2:stim trace for c1
#c3:model schematic
#d:V-V CC trajectories
#e1:spike rasters, V1-V2 (20-100 Hz), LFP (for model_version3)
#e2:stim trace for e1
#e3:model schematic
#f:V-V CC trajectories
#widths and heights of all subfigures IN INCHES
width_traj = 1.2
width_traces = width_fig - 2 * padding - 1.5 * interpadding - width_traj
height_stim = 0.1
height_traces = (1. / 3.) * (height_fig - 2 * padding - 3 * interpadding -
3 * height_stim)
height_traj = (3. / 5.) * (height_traces + height_stim - interpadding)
height_schematic = (2. / 5.) * (height_traces + height_stim - interpadding)
width_schematic = width_traj
#x and y positions of all subfigures and labels IN INCHES
x_a1 = padding
y_a1 = height_fig - padding - height_traces
x_a2 = x_a1
y_a2 = y_a1 - height_stim
x_a3 = width_fig - padding - width_schematic
x_b = width_fig - padding - width_traj
y_b = y_a2
y_a3 = y_b + height_traj + 0.35 * interpadding
x_c1 = x_a1
y_c1 = y_a2 - 1.5 * interpadding - height_traces
x_c2 = x_c1
y_c2 = y_c1 - height_stim
x_c3 = x_a3
x_d = x_b
y_d = y_c2
y_c3 = y_d + height_traj + 0.35 * interpadding
x_e1 = x_a1
y_e1 = padding + height_stim
x_e2 = x_e1
y_e2 = padding
x_e3 = x_a3
y_e3 = padding + height_traj + 0.35 * interpadding
x_f = x_e3
y_f = padding
x_label_a = x_a1 - 0.3
y_label_a = y_a1 + height_traces - 0.25 * interpadding
x_label_b = x_b - 0.4
y_label_b = y_b + height_traj
x_label_c = x_c1 - 0.3
y_label_c = y_c1 + height_traces - 0.25 * interpadding
x_label_d = x_label_b
y_label_d = y_d + height_traj
x_label_e = x_e1 - 0.3
y_label_e = y_e1 + height_traces - 0.25 * interpadding
x_label_f = x_label_b
y_label_f = y_f + height_traj
#x and y positions of all subfigures and labels IN FRACTIONS OF FIG SIZE
x_a1 = (x_a1 / width_fig)
y_a1 = (y_a1 / height_fig)
x_a2 = (x_a2 / width_fig)
y_a2 = (y_a2 / height_fig)
x_a3 = (x_a3 / width_fig)
y_a3 = (y_a3 / height_fig)
x_b = (x_b / width_fig)
y_b = (y_b / height_fig)
x_c1 = (x_c1 / width_fig)
y_c1 = (y_c1 / height_fig)
x_c2 = (x_c2 / width_fig)
y_c2 = (y_c2 / height_fig)
x_c3 = (x_c3 / width_fig)
y_c3 = (y_c3 / height_fig)
x_d = (x_d / width_fig)
y_d = (y_d / height_fig)
x_e1 = (x_e1 / width_fig)
y_e1 = (y_e1 / height_fig)
x_e2 = (x_e2 / width_fig)
y_e2 = (y_e2 / height_fig)
x_e3 = (x_e3 / width_fig)
y_e3 = (y_e3 / height_fig)
x_f = (x_f / width_fig)
y_f = (y_f / height_fig)
x_label_a = (x_label_a / width_fig)
y_label_a = (y_label_a / height_fig)
x_label_b = (x_label_b / width_fig)
y_label_b = (y_label_b / height_fig)
x_label_c = (x_label_c / width_fig)
y_label_c = (y_label_c / height_fig)
x_label_d = (x_label_d / width_fig)
y_label_d = (y_label_d / height_fig)
x_label_e = (x_label_e / width_fig)
y_label_e = (y_label_e / height_fig)
x_label_f = (x_label_f / width_fig)
y_label_f = (y_label_f / height_fig)
#widths and heights of all subfigures IN FRACTIONS OF FIG SIZE
width_traces = (width_traces / width_fig)
height_traces = (height_traces / height_fig)
width_traj = (width_traj / width_fig)
height_traj = (height_traj / height_fig)
width_schematic = (width_schematic / width_fig)
height_schematic = (height_schematic / height_fig)
height_stim = (height_stim / height_fig)
V_mult_factor = 10000.0
LFP_mult_factor = -2.5
fig = plt.figure(figsize=(width_fig, height_fig), dpi=300)
###################### Panel A1: spikes, Vs, LFP
width = width_traces
height = height_traces
x_pos = x_a1
y_pos = y_a1
rect = (x_pos, y_pos, width, height)
fig.text(x_label_a, y_label_a, 'A', fontproperties=fontl)
ax1 = fig.add_axes(rect)
if showfig == True or savefig == True:
plot_model_traces(
ax1, fonts, node_group, N_E, N_I, tau_Em, tau_Im, p_exc, p_inh,
tau_ref_E, tau_ref_I, V_leak_E_min, V_leak_I_min, V_th, V_reset,
V_peak, V_syn_E, V_syn_I, C_E, C_I, g_AE1, g_AI, g_GE, g_GI,
g_extE, g_extI1, g_leak_E, g_leak_I, tau_Fl, tau_AIl, tau_AIr,
tau_AId, tau_AEl, tau_AEr, tau_AEd, tauR_intra, tauD_intra,
tauR_extra, tauD_extra, rateI, rateE, stim_factor,
num_nodes_to_save, num_LFPs, windows, gaps, padding_for_traces,
stim_type, time_to_max, exc_connectivity1, inh_connectivity,
exc_syn_weight_dist, net_state1, ext_SD, int_SD, generate_LFPs,
save_results, num_trials, crit_freq_V, filt_kind_V, crit_freq_LFP,
filt_kind_LFP, V_mult_factor, LFP_mult_factor, run_new_net_sim,
master_folder_path)
###################### Panel A2: stim trace
width = width_traces
height = height_stim
x_pos = x_a2
y_pos = y_a2
rect = (x_pos, y_pos, width, height)
ax2 = fig.add_axes(rect)
plot_stim_trace(ax2, windows, gaps, 0)
ax2.set_xlim(0, sum(windows) + sum(gaps))
###################### Panel A3: model_schematic
width = width_schematic
height = height_schematic
x_pos = x_a3
y_pos = y_a3
rect = (x_pos, y_pos, width, height)
ax3 = fig.add_axes(rect)
if showfig == True or savefig == True:
image_path = 'file:\\' + master_folder_path + \
'\\figures\\random_schematic.png'
image = urllib2.urlopen(image_path)
array = pylab.imread(image)
ax3.imshow(array)
ax3.set_axis_off()
ax3.text(0.5,
1.15,
'random asynchronous',
fontsize=8,
horizontalalignment='center',
verticalalignment='bottom',
transform=ax3.transAxes)
###################### Panel B: CC trajectories
width = width_traj
height = height_traj
x_pos = x_b
y_pos = y_b
rect = (x_pos, y_pos, width, height)
ax4 = fig.add_axes(rect)
fig.text(x_label_b, y_label_b, 'B', fontproperties=fontl)
plot_CC(ax4, fonts, node_group, N_E, N_I, tau_Em, tau_Im, p_exc, p_inh,
tau_ref_E, tau_ref_I, V_leak_E_min, V_leak_I_min, V_th, V_reset,
V_peak, V_syn_E, V_syn_I, C_E, C_I, g_AE1, g_AI, g_GE, g_GI,
g_extE, g_extI1, g_leak_E, g_leak_I, tau_Fl, tau_AIl, tau_AIr,
tau_AId, tau_AEl, tau_AEr, tau_AEd, tauR_intra, tauD_intra,
tauR_extra, tauD_extra, rateI, rateE, stim_factor,
num_nodes_to_save, num_pairs, num_LFPs, windows, gaps,
padding_for_traces, stim_type, time_to_max, exc_connectivity1,
inh_connectivity, exc_syn_weight_dist, net_state1, ext_SD, int_SD,
generate_LFPs, save_results, num_trials, crit_freq_V, filt_kind_V,
run_new_net_sim, get_new_CC_results_all_pairs, master_folder_path)
###################### Panel C1: spikes, Vs, LFP
width = width_traces
height = height_traces
x_pos = x_c1
y_pos = y_c1
rect = (x_pos, y_pos, width, height)
fig.text(x_label_c, y_label_c, 'C', fontproperties=fontl)
ax5 = fig.add_axes(rect)
if showfig == True or savefig == True:
plot_model_traces(
ax5, fonts, node_group, N_E, N_I, tau_Em, tau_Im, p_exc, p_inh,
tau_ref_E, tau_ref_I, V_leak_E_min, V_leak_I_min, V_th, V_reset,
V_peak, V_syn_E, V_syn_I, C_E, C_I, g_AE2, g_AI, g_GE, g_GI,
g_extE, g_extI2, g_leak_E, g_leak_I, tau_Fl, tau_AIl, tau_AIr,
tau_AId, tau_AEl, tau_AEr, tau_AEd, tauR_intra, tauD_intra,
tauR_extra, tauD_extra, rateI, rateE, stim_factor,
num_nodes_to_save, num_LFPs, windows, gaps, padding_for_traces,
stim_type, time_to_max, exc_connectivity2, inh_connectivity,
exc_syn_weight_dist, net_state2, ext_SD, int_SD, generate_LFPs,
save_results, num_trials, crit_freq_V, filt_kind_V, crit_freq_LFP,
filt_kind_LFP, V_mult_factor, LFP_mult_factor, run_new_net_sim,
master_folder_path)
###################### Panel C2: stim trace
width = width_traces
height = height_stim
x_pos = x_c2
y_pos = y_c2
rect = (x_pos, y_pos, width, height)
ax6 = fig.add_axes(rect)
plot_stim_trace(ax6, windows, gaps, 0)
ax6.set_xlim(0, sum(windows) + sum(gaps))
###################### Panel C3: model_schematic
width = width_schematic
height = height_schematic
x_pos = x_c3
y_pos = y_c3
rect = (x_pos, y_pos, width, height)
ax7 = fig.add_axes(rect)
if showfig == True or savefig == True:
image_path = 'file:\\' + master_folder_path + \
'\\figures\\random_schematic.png'
image = urllib2.urlopen(image_path)
array = pylab.imread(image)
ax7.imshow(array)
ax7.set_axis_off()
ax7.text(0.5,
1.15,
'random synchronous',
fontsize=8,
horizontalalignment='center',
verticalalignment='bottom',
transform=ax7.transAxes)
###################### Panel D: CC trajectories
width = width_traj
height = height_traj
x_pos = x_d
y_pos = y_d
rect = (x_pos, y_pos, width, height)
ax8 = fig.add_axes(rect)
fig.text(x_label_d, y_label_d, 'D', fontproperties=fontl)
plot_CC(ax8, fonts, node_group, N_E, N_I, tau_Em, tau_Im, p_exc, p_inh,
tau_ref_E, tau_ref_I, V_leak_E_min, V_leak_I_min, V_th, V_reset,
V_peak, V_syn_E, V_syn_I, C_E, C_I, g_AE2, g_AI, g_GE, g_GI,
g_extE, g_extI2, g_leak_E, g_leak_I, tau_Fl, tau_AIl, tau_AIr,
tau_AId, tau_AEl, tau_AEr, tau_AEd, tauR_intra, tauD_intra,
tauR_extra, tauD_extra, rateI, rateE, stim_factor,
num_nodes_to_save, num_pairs, num_LFPs, windows, gaps,
padding_for_traces, stim_type, time_to_max, exc_connectivity2,
inh_connectivity, exc_syn_weight_dist, net_state2, ext_SD, int_SD,
generate_LFPs, save_results, num_trials, crit_freq_V, filt_kind_V,
run_new_net_sim, get_new_CC_results_all_pairs, master_folder_path)
###################### Panel E1: spikes, Vs, LFP
width = width_traces
height = height_traces
x_pos = x_e1
y_pos = y_e1
rect = (x_pos, y_pos, width, height)
fig.text(x_label_e, y_label_e, 'E', fontproperties=fontl)
ax9 = fig.add_axes(rect)
rateI = 50
rateE = 50
stim_factor = 6
if showfig == True or savefig == True:
plot_model_traces(
ax9, fonts, node_group, N_E, N_I, tau_Em, tau_Im, p_exc, p_inh,
tau_ref_E, tau_ref_I, V_leak_E_min, V_leak_I_min, V_th, V_reset,
V_peak, V_syn_E, V_syn_I, C_E, C_I, g_AE3, g_AI, g_GE, g_GI,
g_extE, g_extI3, g_leak_E, g_leak_I, tau_Fl, tau_AIl, tau_AIr,
tau_AId, tau_AEl, tau_AEr, tau_AEd, tauR_intra, tauD_intra,
tauR_extra, tauD_extra, rateI, rateE, stim_factor,
num_nodes_to_save, num_LFPs, windows, gaps, padding_for_traces,
stim_type, time_to_max, exc_connectivity3, inh_connectivity,
exc_syn_weight_dist, net_state3, ext_SD, int_SD, generate_LFPs,
save_results, num_trials, crit_freq_V, filt_kind_V, crit_freq_LFP,
filt_kind_LFP, V_mult_factor, LFP_mult_factor, run_new_net_sim,
master_folder_path)
###################### Panel E2: stim trace
width = width_traces
height = height_stim
x_pos = x_e2
y_pos = y_e2
rect = (x_pos, y_pos, width, height)
ax10 = fig.add_axes(rect)
plot_stim_trace(ax10, windows, gaps, 0)
ax10.set_xlim(0, sum(windows) + sum(gaps))
###################### Panel E3: model_schematic
width = width_schematic
height = height_schematic
x_pos = x_e3
y_pos = y_e3
rect = (x_pos, y_pos, width, height)
ax11 = fig.add_axes(rect)
if showfig == True or savefig == True:
image_path = 'file:\\' + master_folder_path + \
'\\figures\\small_world_schematic.png'
image = urllib2.urlopen(image_path)
array = pylab.imread(image)
ax11.imshow(array)
ax11.set_axis_off()
ax11.text(0.5,
1.15,
'clustered',
fontsize=8,
horizontalalignment='center',
verticalalignment='bottom',
transform=ax11.transAxes)
###################### Panel F: CC trajectories
width = width_traj
height = height_traj
x_pos = x_f
y_pos = y_f
rect = (x_pos, y_pos, width, height)
ax12 = fig.add_axes(rect)
fig.text(x_label_f, y_label_f, 'F', fontproperties=fontl)
plot_CC(ax12, fonts, node_group, N_E, N_I, tau_Em, tau_Im, p_exc, p_inh,
tau_ref_E, tau_ref_I, V_leak_E_min, V_leak_I_min, V_th, V_reset,
V_peak, V_syn_E, V_syn_I, C_E, C_I, g_AE3, g_AI, g_GE, g_GI,
g_extE, g_extI3, g_leak_E, g_leak_I, tau_Fl, tau_AIl, tau_AIr,
tau_AId, tau_AEl, tau_AEr, tau_AEd, tauR_intra, tauD_intra,
tauR_extra, tauD_extra, rateI, rateE, stim_factor,
num_nodes_to_save, num_pairs, num_LFPs, windows, gaps,
padding_for_traces, stim_type, time_to_max, exc_connectivity3,
inh_connectivity, exc_syn_weight_dist, net_state3, ext_SD, int_SD,
generate_LFPs, save_results, num_trials, crit_freq_V, filt_kind_V,
run_new_net_sim, get_new_CC_results_all_pairs, master_folder_path)
###################### Calculate and report CC values for clustered
###################### network with no synaptic depression
ext_SD = False
int_SD = False
report_CC(
node_group, N_E, N_I, tau_Em, tau_Im, p_exc, p_inh, tau_ref_E,
tau_ref_I, V_leak_E_min, V_leak_I_min, V_th, V_reset, V_peak, V_syn_E,
V_syn_I, C_E, C_I, g_AE3, g_AI, g_GE, g_GI, g_extE, g_extI3, g_leak_E,
g_leak_I, tau_Fl, tau_AIl, tau_AIr, tau_AId, tau_AEl, tau_AEr, tau_AEd,
tauR_intra, tauD_intra, tauR_extra, tauD_extra, rateI, rateE,
stim_factor, num_nodes_to_save, num_pairs, num_LFPs, windows, gaps,
padding_for_traces, stim_type, time_to_max, exc_connectivity3,
inh_connectivity, exc_syn_weight_dist, net_state3, ext_SD, int_SD,
generate_LFPs, save_results, num_trials, crit_freq_V, filt_kind_V,
run_new_net_sim, get_new_CC_results_all_pairs, master_folder_path)
if savefig == True:
figpath = master_folder_path + '\\figures\\Fig5'
fig.savefig(figpath + '.png', dpi=300)
figpath2 = '\\home\\caleb\\Dropbox\\Fig5'
fig.savefig(figpath2 + '.png', dpi=300)
if showfig == True:
pylab.show()
import sys import pymaxflow import pylab import numpy as np eps = 0.01 im = pylab.imread(sys.argv[1]).astype(np.float32) indices = np.arange(im.size).reshape(im.shape).astype(np.int32) g = pymaxflow.PyGraph(im.size, im.size * 3) g.add_node(im.size) # adjacent diffs = np.abs(im[:, 1:] - im[:, :-1]).ravel() + eps e1 = indices[:, :-1].ravel() e2 = indices[:, 1:].ravel() g.add_edge_vectorized(e1, e2, diffs, 0 * diffs) # adjacent up diffs = np.abs(im[1:, 1:] - im[:-1, :-1]).ravel() + eps e1 = indices[1:, :-1].ravel() e2 = indices[:-1, 1:].ravel() g.add_edge_vectorized(e1, e2, diffs, 0 * diffs) # adjacent down diffs = np.abs(im[:-1, 1:] - im[1:, :-1]).ravel() + eps e1 = indices[:-1, :-1].flatten() e2 = indices[1:, 1:].ravel() g.add_edge_vectorized(e1, e2, diffs, 0 * diffs)
zorder=5,
clip_on=False,
transform=fig.transFigure)
for lx, ly in zip(linex, liney):
ax.plot(lx, ly, **kwargs)
kwargs = dict(fontsize=6,
va="baseline",
ha="center",
color=obscolor["computational"],
transform=fig.transFigure)
ax.text(l1 + wi / 2, bi + 0.1415, observernames["computational"], **kwargs)
# # # PLOT OBSERVER MODEL SKETCHES # # #
if "A" in PLOT:
ax = axes["observers"]
ax.imshow(pl.imread(f"./panel/observermodelspictogram_wide.png"))
ax.set_frame_on(False)
print_panel_label("A", ax, abs_x=0.025, abs_y=0.97)
# # # PLOT ALT. OBSERVERS' PERFORMANCES # # #
if "B" in PLOT:
print_panel_label("B", ax, abs_x=0.025, abs_y=0.795)
# PERPARE GENERAL VARS
pc = perf_level["chance"]
pp = perf_level["perfect"]
nObs = len(observers)
nCond = len(conds)
nTracker = len(visible_trackers)
ymin, ymax = 10., 0.
from scipy.stats import sem
ax = axes["bars"]
import matplotlib.pyplot as plt import numpy as np import ocr from pylab import imread, imshow, figure, show, subplot, plot, scatter from scipy.cluster.vq import kmeans, vq from skimage import data, img_as_uint, img_as_float from skimage.external.tifffile import imsave from skimage.filters import threshold_otsu, threshold_adaptive, threshold_yen from skimage.segmentation import clear_border imageFile = '../pics/1.png' image = imread(imageFile) img = data.imread(imageFile, as_grey=True) global_thresh = threshold_yen(img) # True False binary matrix represent color value of the img using global thresholding binary_global = img > global_thresh block_size = 40 # True False binary matrix represent color value of the img using adaptive thresholding binary_adaptive = threshold_adaptive(img, block_size, offset=0) # 0 1 binary matrix img_bin_global = clear_border(img_as_uint(binary_global)) # 0 1 binary matrix img_bin_adaptive = clear_border(img_as_uint(binary_adaptive)) savedImg = img_as_float(binary_adaptive)
myFiles = [] # 文件列表
for (root, dirs, files) in os.walk(input):
for file in files:
# 如果【不是标注图】并且【是jpg格式】
if file.find("_gt") < 0 and file.find('.jpg') >= 0:
myFiles.append(file) # 添加到文件列表
break # os.walk只对当前目录遍历一层
# 对文件列表进行洗牌
random.shuffle(myFiles)
length = len(myFiles) * 4 # 因为需要旋转4个角度
# 建立一个编号列表。线性递增表为[0,1,...,length-1]，对这个线性递增表进行洗牌，作为输出文件的序号
lst = build_shuffle_list(length)
lst = list(lst)
for i, s in enumerate(myFiles): #i：操作序号，s：当前文件名
# 分离文件名和扩展名。name为当前文件文件名（无后缀）
name = os.path.splitext(s)[0]
print(i, name)
imgName = f'{input}/' + name + '.jpg' #输入原图
dataName = f'{input}/' + name + '_gt.jpg' #输入标注图
img = plt.imread(imgName) #图片读取
data = plt.imread(dataName)
# 直接保存
saveFiles(img, data, lst)
for j in range(3): #连续旋转3次
img = RotateClockWise90(img)
data = RotateClockWise90(data)
saveFiles(img, data, lst)
add_log('结束时间：')
# 对于长操作，操作后关机
# os.system('poweroff')
print(fn)
try:
os.makedirs(os.path.dirname(fn) +
'/../training_images') #, exists_ok = True)
except:
pass
try:
os.makedirs(os.path.dirname(fn) +
'/../test_images') #, exists_ok = True)
except:
pass
#raw = rawpy.imread(fn)
#bayer = raw.raw_image
bayer = pylab.imread(fn)
pylab.imshow(bayer)
print(bayer.shape)
for i in range(xmin, xmax, width):
for j in range(ymin, ymax, height):
try:
if i > j:
train_01 = bayer[j:j + height, i:i + width]
if train_01.shape[0] > 0 and train_01.shape[1] > 0:
train_01_fn = fn.replace(
'raw_images', 'training_images').replace(
'.jpg', '-train_01_%d_%d.jpg' % (i, j))
SCALE = 1.5
if event.button == "up":
width = (x2 - x1) / SCALE
height = (y2 - y1) / SCALE
elif event.button == "down":
width = (x2 - x1) * SCALE
height = (y2 - y1) * SCALE
ax.set_xlim(x - width * xp, x + width * (1 - xp))
ax.set_ylim(y - height * yp, y + height * (1 - yp))
pylab.draw()
fig.canvas.mpl_connect('scroll_event', on_scroll)
#Show background image
bg = pylab.imread(configuration["bg_picture"])
pylab.subplots_adjust(left=0.05, right=0.95, top=0.95, bottom=0.05)
ax = pylab.subplot(111)
pylab.imshow(bg, extent=configuration["extent"], origin='lower')
'''
#Make the plots
labels = marks.keys()
import mission.layout.layout as layout
desired_labels = set(layout.places)
removed_labels = set(labels).difference(desired_labels)
if removed_labels:
print "The following labels no longer appear in layout.places"
print ", ".join(x for x in removed_labels)
resp = raw_input("Would you like to remove them from the layout? y/N")
fresize.restype = None
fresize.argtypes = [
ctypeslib.ndpointer(dtype=datatype, ndim=3), c_int,
ctypeslib.ndpointer(dtype=datatype, ndim=3), c_int, c_int, c_int
]
ddims = [
int(round(sdims[0] * scale)),
int(round(sdims[1] * scale)), sdims[2]
]
mxdst = zeros((ddims), dtype=datatype)
tmp = zeros((ddims[0], sdims[1], sdims[2]), dtype=datatype)
img1 = img
t1 = time()
fresize(img1, sdims[0], tmp, ddims[0], sdims[1], sdims[2])
fresize(tmp, sdims[1], mxdst, ddims[1], ddims[0], sdims[2])
t2 = time()
return mxdst.reshape(ddims[2], ddims[1], ddims[0]).T
if __name__ == "__main__":
from numpy.random import random_integers
from time import time
from pylab import imread, figure, imshow
from ctypes import c_float, c_double, c_int
img = imread("test.png").astype(c_double)
imshow(img)
img1 = resize(img, 0.25)
figure()
imshow(img1)
###Cargar Imagen original###
import matplotlib.pyplot as plt
import numpy as np
from pylab import imread, imshow
img = imread('D:\Esteban\Desktop\Ezatara\EZ\EZ-Images\grim.jpg').astype(np.float32) #Se lee archivo con la funcion imread de pylab
image = img / 255
imshow(image) #Se muestra la imagen con la funcion ishow de la libreria matplotlib
#Cargar y Leer CSV
import numpy as np
data = np.genfromtxt('D:/Esteban/Desktop/Ezatara/EZ-U/Aseguramiento/POC/train.csv', delimiter = ',', skip_header=1, dtype=None) #Se utiliza numpy y su funcion
#de genfromtxt para leer el archivo
for row in data:
print(row) #Se imprimen las row del archivo en forma de array
#print(row[0], row[1], row[2]) #Para cargar el csv pero no desplegar de forma de array
###Cargar la imagen y ponerle y manejar el contraste a la misma###
from scipy import ndimage
import matplotlib.pyplot as plt
import numpy as np
from pylab import imread, imshow, gray, mean
img = imread('D:\Esteban\Desktop\Ezatara\EZ\EZ-Images\grim.jpg').astype(np.float32) #Se lle el archivo con imread de pylab
dimshow(resampled_img)
plt.subplot(1, 2, 2)
dimshow(resampled_mask)
ps.savefig()
# feed it to "enhance"
enhance.update(resampled_mask.ravel(), img[Yi, Xi, :])
plt.clf()
dimshow(enhance.enhI.reshape((H, W, 3)))
ps.savefig()
sys.exit(0)
# test stretching
img = plt.imread('demo/apod1.jpg')
(H, W, B) = img.shape
print('Image', img.shape, img.dtype)
imx = np.sqrt(img.astype(np.float32) / 255.)
enhance = EnhanceImage(H * W, B)
enhance.update(np.ones((H * W), bool), imx.reshape((-1, B)))
en = enhance.enhI.reshape((H, W, B))
plt.clf()
plt.imshow(en, interpolation='nearest', origin='lower')
plt.savefig('en.png')
stretch = enhance.stretch_to_match(img).reshape((H, W, B))
from __future__ import division import glob import os.path import numpy as np import amitgroup as ag import pylab as plt from config import SETTINGS files = glob.glob(os.path.join(SETTINGS['src_dir'], '*.png')) im = plt.imread(files[32]) imgrey = im[...,:3].mean(axis=2).astype(np.float64) edges = ag.features.bedges(imgrey, radius=1, firstaxis=True) edges2 = ag.features.bedges_from_image(im.astype(np.float64), radius=1, firstaxis=True) num_edges, num_edges2 = edges.sum(), edges2.sum() print "Edges (grayscale):", num_edges print "Edges (RGB):", num_edges2 print "Increase:", num_edges2/num_edges ag.plot.images(np.r_[edges, edges2])
import os
sys.path.append('../../')
from Utils.StainNormalization import StainNormalizationWithVector,StainNormalizationWithVector2
h_ref = np.array([[0.24502717, 0.80708244, 0.34289746],[0.63904356, 0.67133489, 0.30034316]])
proot ='/home/zyx31/DATA_CRLM/Patches/Patches_Level0/Patches_1024/All/'
proot_norm = '/home/zyx31/DATA_CRLM/Patches/Patches_Level0/Patches_1024/Norm/'
pindex = 0
plist = os.listdir(proot)
for pindex in range(len(plist)):
ppath = proot + plist[pindex]
#ppath = proot + fname_dict['T'][0]
print("processing %s"%plist[pindex])
try:
patch = np.array(plt.imread(ppath))[:,:,:3]
normp = StainNormalizationWithVector(patch,h_ref=h_ref,\
illuminant_ref=[255,255,255],no_channel=2,Df=5, init=[260,320])
#plt.imshow(normp)
plt.imsave(proot_norm+plist[pindex],normp)
except Exception as e:
print(e)
if pindex %200 ==0:
print("----------------->",pindex)
return AttentionModule.aggregate(outputs, gates, last_output,
last_gate), reg_loss
if __name__ == '__main__':
import time
t = time.time()
net = WideResNetAttention(1000,
attention_depth=3,
nheads=4,
has_gates=True,
pretrain_path="pretrained/wrn-50-2.t7").eval()
print(time.time() - t)
import pylab
import cv2
import numpy as np
im = pylab.imread('./demo/goldfish.jpeg') / 255.
im = cv2.resize(im, (224, 224))
im = im.transpose(2, 0, 1)
im -= np.array([0.485, 0.456, 0.406]).reshape((3, 1, 1))
im /= np.array([0.229, 0.224, 0.225]).reshape((3, 1, 1))
im = torch.Tensor(im[None, ...])
t = time.time()
out = net(Variable(im))
print(time.time() - t)
max, ind = torch.max(out, 1)
# max, ind = torch.max(net.linear(out), 1)
# print(max)
print(ind)
from pylab import imread, imshow, figure, show, subplot
import matplotlib.pyplot as plt
from numpy import reshape, uint8, flipud
from sklearn.cluster import KMeans
from copy import deepcopy
img = imread('sample.jpeg')
pixel = reshape(img, (img.shape[0] * img.shape[1],
3)) # picture can change from 8 * 8 * 3 into 64 * 3
pixel_new = deepcopy(pixel)
print(img.shape)
print(pixel_new.shape)
model = KMeans(n_clusters=5)
labels = model.fit_predict(pixel)
palette = model.cluster_centers_ # compress the color, chage the matrix of the picture all use center point
for i in range(len(pixel)):
pixel_new[i, :] = palette[
labels[i]] # use label as subscript to chagne every matrix context
imshow(reshape(pixel_new, (img.shape[0], img.shape[1], 3)))
plt.show()
def plotarea(ra, dec, radius, name, prefix, tims=None, rds=[]):
from astrometry.util.util import Tan
W, H = 512, 512
scale = (radius * 60. * 4) / float(W)
print 'SDSS jpeg scale', scale
imgfn = 'sdss-mosaic-%s.png' % prefix
if not os.path.exists(imgfn):
url = (('http://skyservice.pha.jhu.edu/DR9/ImgCutout/getjpeg.aspx?' +
'ra=%f&dec=%f&scale=%f&width=%i&height=%i') %
(ra, dec, scale, W, H))
f = urllib2.urlopen(url)
of, tmpfn = tempfile.mkstemp(suffix='.jpg')
os.close(of)
of = open(tmpfn, 'wb')
of.write(f.read())
of.close()
cmd = 'jpegtopnm %s | pnmtopng > %s' % (tmpfn, imgfn)
os.system(cmd)
# Create WCS header for it
cd = scale / 3600.
args = (ra, dec, W / 2. + 0.5, H / 2. + 0.5, -cd, 0., 0., -cd, W, H)
wcs = Tan(*[float(x) for x in args])
plt.clf()
I = plt.imread(imgfn)
plt.imshow(I, interpolation='nearest', origin='lower')
x, y = wcs.radec2pixelxy(ra, dec)
R = radius * 60. / scale
ax = plt.axis()
plt.gca().add_artist(
matplotlib.patches.Circle(xy=(x, y),
radius=R,
color='g',
lw=3,
alpha=0.5,
fc='none'))
if tims is not None:
print 'Plotting outlines of', len(tims), 'images'
for tim in tims:
H, W = tim.shape
twcs = tim.getWcs()
px, py = [], []
for x, y in [(1, 1), (W, 1), (W, H), (1, H), (1, 1)]:
rd = twcs.pixelToPosition(x, y)
xx, yy = wcs.radec2pixelxy(rd.ra, rd.dec)
print 'x,y', x, y
x1, y1 = twcs.positionToPixel(rd)
print ' x1,y1', x1, y1
print ' r,d', rd.ra, rd.dec,
print ' xx,yy', xx, yy
px.append(xx)
py.append(yy)
plt.plot(px, py, 'g-', lw=3, alpha=0.5)
# plot full-frame image outline too
# px,py = [],[]
# W,H = 2048,1489
# for x,y in [(1,1),(W,1),(W,H),(1,H),(1,1)]:
# r,d = twcs.pixelToRaDec(x,y)
# xx,yy = wcs.radec2pixelxy(r,d)
# px.append(xx)
# py.append(yy)
# plt.plot(px, py, 'g-', lw=1, alpha=1.)
if rds is not None:
px, py = [], []
for ra, dec in rds:
print 'ra,dec', ra, dec
xx, yy = wcs.radec2pixelxy(ra, dec)
px.append(xx)
py.append(yy)
plt.plot(px, py, 'go')
plt.axis(ax)
fn = '%s.png' % prefix
plt.savefig(fn)
print 'saved', fn
# with slim.arg_scope(vgg.vgg_arg_scope()):
# output, _ = vgg.vgg_16(normalized_image, num_classes = 1000, is_training = False)
# probabilities = tf.nn.softmax(output)
# init_fn = slim.assign_from_checkpoint_fn(os.path.join(checkpoints_dir, 'vgg_16.ckpt'), slim.get_model_variables('vgg_16'))
# sess = tf.Session()
# init_fn(sess)
# pass
if __name__ == "__main__":
# Set some parameters
image_size = vgg.vgg_16.default_image_size
batch_size = 1
image_file = "./data/imagenet/catdog/catdog.jpg"
image = plt.imread(image_file)
# plt.imshow(image)
# plt.annotate('Something', xy = (0.05, 0.95), xycoords = 'axes fraction')
# plt.show()
labels = imagenet.create_readable_names_for_imagenet_labels()
# Define graph
with tf.Graph().as_default():
x = tf.placeholder(dtype=tf.float32, shape=(image_size, image_size, 3))
normalized_image = vgg_preprocessing.preprocess_image(
x, image_size, image_size, is_training=False)
normalized_images = tf.expand_dims(normalized_image, 0)
with slim.arg_scope(vgg.vgg_arg_scope()):
output, _ = vgg.vgg_16(normalized_images,
num_classes=1000,
xp, yp = (x - x1) / (x2 - x1), (y - y1) / (y2 - y1)
SCALE = 1.5
if event.button == "up":
width = (x2 - x1) / SCALE
height = (y2 - y1) / SCALE
elif event.button == "down":
width = (x2 - x1) * SCALE
height = (y2 - y1) * SCALE
ax.set_xlim(x - width * xp, x + width * (1 - xp))
ax.set_ylim(y - height * yp, y + height * (1 - yp))
pylab.draw()
fig.canvas.mpl_connect('scroll_event', on_scroll)
bg = pylab.imread(
os.path.join(os.path.join(path, 'images'), args.locale + '.jpg'))
pylab.subplots_adjust(left=0.05, right=0.95, top=0.95, bottom=0.05)
ax = pylab.subplot(111)
bg_width = 27.7
bg_height = 16.66
corner_offset_x = -5.5
corner_offset_y = -bg_height + 3
extent = (corner_offset_x, bg_width + corner_offset_x, corner_offset_y,
bg_height + corner_offset_y)
pylab.imshow(bg, extent=extent, origin='lower')
xs = [m['position'][1] for m in marks]
ys = [m['position'][0] for m in marks]
reverse = {
def get_classifier_params(self):
return self.linear.parameters()
def forward(self, x):
x = self.conv1(x)
x = self.bn1(x)
x = F.relu(x)
x = F.max_pool2d(x, 3, 2, 1)
x = self.group0(x)
x = self.group1(x)
x = self.group2(x)
x = self.group3(x)
return F.log_softmax(self.linear(x.mean(3).mean(2)), dim=1)
if __name__ == '__main__':
net = WideResNet().eval()
import pylab
import cv2
import numpy as np
cat = pylab.imread('demo/goldfish.jpeg') / 255.
cat = cv2.resize(cat, (224, 224))
cat = cat.transpose(2, 0, 1)
cat -= np.array([0.485, 0.456, 0.406]).reshape((3, 1, 1))
cat /= np.array([0.229, 0.224, 0.225]).reshape((3, 1, 1))
cat = torch.Tensor(cat[None, ...])
out = net(Variable(cat))
max, ind = torch.max(net.linear(out), 1)
# print(max)
print(ind)
from scipy.misc import lena
from pylab import imread
from scipy.ndimage import gaussian_filter
from stl_tools import numpy2stl, text2png
"""
Some quick examples
"""
A = lena() # load Lena image, shrink in half
A = gaussian_filter(A, 1) # smoothing
numpy2stl(A, "examples/Lena.stl", scale=0.1, solid=False)
A = 256 * imread("examples/example_data/NASA.png")
A = A[:, :, 2] + 1.0 * A[:, :, 0] # Compose RGBA channels to give depth
A = gaussian_filter(A, 1) # smoothing
numpy2stl(A, "examples/NASA.stl", scale=0.05, mask_val=5., solid=True)
A = 256 * imread("examples/example_data/openmdao.png")
A = A[:, :,
0] + 1. * A[:, :, 3] # Compose some elements from RGBA to give depth
A = gaussian_filter(A, 2) # smoothing
numpy2stl(A,
"examples/OpenMDAO-logo.stl",
scale=0.05,
mask_val=1.,
min_thickness_percent=0.005,
solid=True)
text = ("$\oint_{\Gamma} (A\, dx + B\, dy) = \iint_{U} \left(\\frac{\partial "
"B}{\partial x} - \\frac{\partial A}{\partial y}\\right)\ dxdy$ \n\n "
"$\\frac{\partial \\rho}{\partial t} + \\frac{\partial}{\partial x_j}"
def render(input_text, fnum=1, dpath=None, verbose=True):
"""
fixme or remove
"""
import pylab as plt
import matplotlib as mpl
#verbose = True
text = make_full_document(input_text)
cwd = os.getcwd()
if dpath is None:
text_dir = join(cwd, 'tmptex')
else:
text_dir = dpath
util_path.ensuredir(text_dir, verbose=verbose)
text_fname = 'latex_formatter_temp.tex'
text_fpath = join(text_dir, text_fname)
pdf_fpath = splitext(text_fpath)[0] + '.pdf'
jpg_fpath = splitext(text_fpath)[0] + '.jpg'
try:
os.chdir(text_dir)
util_io.write_to(text_fpath, text)
pdflatex_args = ('pdflatex', '-shell-escape', '--synctex=-1',
'-src-specials', '-interaction=nonstopmode')
args = pdflatex_args + (text_fpath, )
util_cplat.cmd(*args, verbose=verbose)
assert util_path.checkpath(pdf_fpath, verbose=verbose), 'latex failed'
# convert latex pdf to jpeg
util_cplat.cmd('convert',
'-density',
'300',
pdf_fpath,
'-quality',
'90',
jpg_fpath,
verbose=verbose)
assert util_path.checkpath(jpg_fpath,
verbose=verbose), 'imgmagick failed'
tex_img = plt.imread(jpg_fpath)
# Crop img bbox
nonwhite_x = np.where(tex_img.flatten() != 255)[0]
nonwhite_rows = nonwhite_x // tex_img.shape[1]
nonwhite_cols = nonwhite_x % tex_img.shape[1]
x1 = nonwhite_cols.min()
y1 = nonwhite_rows.min()
x2 = nonwhite_cols.max()
y2 = nonwhite_rows.max()
#util.embed()
cropped = tex_img[y1:y2, x1:x2]
fig = plt.figure(fnum)
fig.clf()
ax = fig.add_subplot(1, 1, 1)
ax.imshow(cropped, cmap=mpl.cm.gray)
#mpl.rc('text', usetex=True)
#mpl.rc('font', family='serif')
#plt.figure()
#plt.text(9, 3.4, text, size=12)
except Exception as ex:
print('LATEX ERROR')
print(text)
print(ex)
print('LATEX ERROR')
pass
finally:
os.chdir(cwd)
def testSubpixelShift(self):
W, H = 100, 100
cx, cy = 34.2, 56.7
X, Y = np.meshgrid(np.arange(W), np.arange(H))
S = 2.
G = 1. / (2. * pi * S**2) * np.exp(-((X - cx)**2 + (Y - cy)**2) /
(2. * S**2))
#G /= G.max()
print G.sum()
print np.sum(G * X), np.sum(G * Y)
imfn = 'test-plotstuff-2.fits'
pyfits.writeto(imfn, G, clobber=True)
wcs = anwcs_create_box(33.5, 10.4, 1., W, H)
anwcs_rotate_wcs(wcs, 25.)
wcsfn = 'test-plotstuff-2.wcs'
anwcs_write(wcs, wcsfn)
plot = Plotstuff()
plot.outformat = PLOTSTUFF_FORMAT_PNG
plot.size = (W, H)
plot.wcs_file = wcsfn
plot.color = 'black'
plot.plot('fill')
im = plot.image
im.image_low = 0
im.image_high = G.max()
plot_image_set_wcs(im, wcsfn, 0)
plot_image_set_filename(im, imfn)
plot.plot('image')
plotfn = 'test-plotstuff-2.png'
plot.write(plotfn)
I = plt.imread(plotfn)
I = I[:, :, 0]
sx, sy = (I * X).sum() / I.sum(), (I * Y).sum() / I.sum()
print sx, sy
ex, ey = cx, cy
self.assertTrue(abs(sx - ex) < 0.1)
self.assertTrue(abs(sy - ey) < 0.1)
# Shift the plot's WCS CRPIX.
dx = 0.25
dy = 0.3
sip = anwcs_get_sip(wcs)
tan = sip.wcstan
plotstuff_set_wcs(plot.pargs, wcs)
xy = plot.xy
plot_xy_set_wcs_filename(xy, wcsfn)
plot.marker = 'crosshair'
# Move the plot WCS origin
for step in range(16):
if step < 8:
tan.set_crpix(tan.crpix[0] + dx, tan.crpix[1])
ex += dx
else:
tan.set_crpix(tan.crpix[0], tan.crpix[1] + dy)
ey += dy
#anwcs_print_stdout(wcs)
plot.color = 'black'
plot.plot('fill')
plot.plot('image')
plotfn = 'test-plotstuff-2b%i.png' % step
plot.write(plotfn)
#anwcs_write(plot.pargs.wcs, 'test-plotstuff-2b.wcs')
I = plt.imread(plotfn)
I = I[:, :, 0]
sx, sy = (I * X).sum() / I.sum(), (I * Y).sum() / I.sum()
print sx, sy
print ex, ey
self.assertTrue(abs(sx - ex) < 0.1)
self.assertTrue(abs(sy - ey) < 0.1)
plot.color = 'red'
plot_xy_clear_list(xy)
# don't plot at sx,sy / ex,ey -- original image coords
# are unchanged.
plot_xy_vals(xy, cx + 1, cy + 1)
plot.plot('xy')
plotfn = 'test-plotstuff-2c%i.png' % step
plot.write(plotfn)
# visual check that plot xy symbols match source centroid -- yes
# Scan image WCS in RA,Dec; check that recovered source position
# through plot WCS matches.
# reset...
plot.wcs_file = wcsfn
plotwcs = anwcs_open(wcsfn, 0)
wcs = anwcs_open(wcsfn, 0)
im.wcs = wcs
sip = anwcs_get_sip(wcs)
tan = sip.wcstan
ddec = 1.2 * 1. / W
ok, era, edec = anwcs_pixelxy2radec(wcs, cx, cy)
print era, edec
for step in range(16):
tan.set_crval(tan.crval[0], tan.crval[1] + ddec)
edec += ddec
plot.color = 'black'
plot.plot('fill')
plot.plot('image')
plotfn = 'test-plotstuff-2d%i.png' % step
plot.write(plotfn)
I = plt.imread(plotfn)
I = I[:, :, 0]
sx, sy = (I * X).sum() / I.sum(), (I * Y).sum() / I.sum()
#print sx,sy
ok, ra, dec = anwcs_pixelxy2radec(plotwcs, sx, sy)
#print era,edec
#print ra,dec
print 'dRA,dDec', ra - era, dec - edec
self.assertTrue(abs(ra - era) < 1e-4)
self.assertTrue(abs(dec - edec) < 1e-4)
import scipy import numpy import pylab from focal import * # load image filename = "./t10k-images-idx3-ubyte__idx_000__lbl_7_.png" img = pylab.imread(filename) #grayscale image pylab.figure() pylab.imshow(img, cmap="Greys_r") fcl = Focal() num_kernels = len(fcl.kernels.full_kernels) # four simulated layers # spikes contains a rank-ordered list of triples with the following # information: # [ pixel/neuron index (int), pixel value (float), layer id (int) ] spikes = fcl.apply(img) # we convert the spike list to 4 images, spike_imgs is a dictionary # containing an image per simulated layer spike_imgs = spike_trains_to_images_g(spikes, img, num_kernels) pylab.figure() i = 1 for k in spike_imgs.keys(): pylab.subplot(2, 2, i) pylab.imshow(spike_imgs[k], cmap="Greys_r") i += 1 #convert to spike source array
# Regularization parameters for kernel estimation
f_alpha = 0. # promotes smoothness
f_beta = 0.1 # Thikhonov regularization
optiter = 500 # number of iterations for minimization
tol = 1e-10 # tolerance for when to stop minimization
# ============================================================================
# Hopefully no need to edit anything below
# ----------------------------------------------------------------------------
# Some more code for backuping the results
# ----------------------------------------------------------------------------
if backup:
# Create helper functions for file handling
yload = lambda i: 1. * pl.imread(FILENAME(i)).astype(np.float32)
# For backup purposes
EXPPATH = '%s/%s_sf%dx%d_csf%dx%d_maxiter%d_alpha%.2f_beta%.2f' % \
(RESPATH,ID,sf[0],sf[1],csf[0],csf[1],optiter,f_alpha,f_beta)
xname = lambda i: '%s/x_%04d.png' % (EXPPATH,i)
yname = lambda i: '%s/y_%04d.png' % (EXPPATH,i)
fname = lambda i: '%s/f_%04d.png' % (EXPPATH,i)
# Create results path if not existing
try:
os.makedirs(EXPPATH)
except:
pass
def Load(filename):
'''
load an image stored in filename are return as a numpy object
'''
return (pyl.imread(filename))
p = np.size(coeff,axis=1)
idx = np.argsort(latent) # sorting the eigenvalues
idx = idx[::-1] # in ascending order
# sorting eigenvectors according to the sorted eigenvalues
coeff = coeff[:,idx]
latent = latent[idx] # sorting eigenvalues
if numpc < p and numpc >= 0:
coeff = coeff[:,range(numpc)] # cutting some PCs if needed
score = np.dot(coeff.T,M) # projection of the data in the new space
return coeff, score, latent
#from pylab import imread,subplot,imshow,title,gray,figure,show,NullLocator
A = pl.imread('dog.jpg') # load an image
A = np.mean(A,2) # to get a 2-D array
full_pc = np.size(A,axis=1) # numbers of all the principal components
i = 1
dist = []
for numpc in range(0,full_pc+10,10):
coeff, score, latent = princomp(A,numpc)
Ar = np.dot(coeff,score).T + np.mean(A,axis=0) # image reconstruction
# difference in Frobenius norm
dist.append(linalg.norm(A-Ar,'fro'))
# showing the pics reconstructed with less than 50 PCs
if numpc <= 25:
ax = pl.subplot(2,3,i,frame_on=False)
ax.xaxis.set_major_locator(pl.NullLocator()) # remove ticks
ax.yaxis.set_major_locator(pl.NullLocator())
i += 1
# ml - ap - dv
ml, ap, dv = location
sliceIndex = 0
for (k, v) in sliceBounds.iteritems():
if ap > v:
sliceIndex = k
break
if sliceIndex == 0:
physio.utils.error("Could not find slice index for: %f" % \
ap, ValueError)
atlasDir = config.get('filesystem', 'atlas')
sliceFile = '%s/%03i.png' % (atlasDir, k)
im = pl.imread(sliceFile)
pl.imshow(im)
# convert ml, dv to pixel coordinates
x, y = skull_to_pixel(ml, -dv, sliceIndex, im.shape)
logging.debug("Channel: %i at %.3f %.3f %.3f" % (ch, ml, ap, dv))
pl.scatter(x, y, color='r')
pl.axhline(y, linestyle='-.', color='b')
pl.axvline(x, linestyle='-.', color='b')
pl.text(x - 10, y + 10, "%.3f, %.3f, %.3f" % (ml, ap, dv), \
ha='right', va='top')
pl.title("Channel: %i" % ch)
pl.xlabel("ML (mm)")
pl.ylabel("DV (mm)")