Skip to content

Commit

Permalink
Issue #53
Browse files Browse the repository at this point in the history
  • Loading branch information
wolkovik committed Jan 27, 2014
1 parent 4edb773 commit 6354ac4
Showing 1 changed file with 53 additions and 99 deletions.
152 changes: 53 additions & 99 deletions src/uiThreshold.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -60,9 +60,9 @@ class uiThreshold:
seeds - matice s kliknutim uzivatele- pokud se maji vracet specifikce objekty
cmap - grey
"""
def __init__(self, data, voxel, threshold = -1, interactivity = True,
number = 100.0, inputSigma = -1, nObj = 10, biggestObjects = True, binaryClosingIterations = 1,
binaryOpeningIterations = 1, seeds = None, cmap = matplotlib.cm.Greys_r):
def __init__(self, data, voxel, threshold=-1, interactivity=True,
number=100.0, inputSigma=-1, nObj=10, biggestObjects=True, binaryClosingIterations=1,
binaryOpeningIterations=1, seeds=None, cmap=matplotlib.cm.Greys_r):

print('Spoustim prahovani dat...')

Expand Down Expand Up @@ -113,7 +113,8 @@ def __init__(self, data, voxel, threshold = -1, interactivity = True,

self.fig = matpyplot.figure()

## Maximalni a minimalni pouzita hodnota prahovani v datech (bud v celych datech nebo vybranych seedu)
## Maximalni a minimalni pouzita hodnota prahovani v datech (bud v
## celych datech nebo vybranych seedu)
self.min0 = numpy.amin(numpy.amin(self.data, axis = 0))
if self.seeds == None:

Expand All @@ -129,25 +130,31 @@ def __init__(self, data, voxel, threshold = -1, interactivity = True,
tmpSeed = 0
dim = numpy.ndim(self.data)
for index in range(0, stop):
## Tady se ukladaji labely na mistech, ve kterych kliknul uzivatel.
## Tady se ukladaji labely na mistech, ve kterych kliknul
## uzivatel.
if dim == 3:
## 3D data.
tmpSeed = self.data[ self.seeds[0][index], self.seeds[1][index], self.seeds[2][index] ]
tmpSeed = self.data[self.seeds[0][index], self.seeds[1][index], self.seeds[2][index]]
elif dim == 2:
## 2D data.
tmpSeed = self.data[ self.seeds[0][index], self.seeds[1][index] ]
tmpSeed = self.data[self.seeds[0][index], self.seeds[1][index]]

## Tady opet pocitam s tim, ze oznaceni nulou pripada cerne oblasti (pozadi).
## Tady opet pocitam s tim, ze oznaceni nulou pripada cerne
## oblasti (pozadi).
if tmpSeed != 0:
## Pokud se nejedna o pozadi (cernou oblast), tak se novy seed ulozi do pole "arrSeed"
## Pokud se nejedna o pozadi (cernou oblast), tak se
## novy seed ulozi do pole "arrSeed"
self.arrSeed.append(tmpSeed)

## Pokud existuji vhodne labely, vytvori se nova data k vraceni.
## Pokud ne, vrati se "None" typ. { Deprecated: Pokud ne, vrati se cela nafiltrovana data, ktera do funkce prisla (nedojde k vraceni specifickych objektu). }
## Pokud existuji vhodne labely, vytvori se nova data k
## vraceni.
## Pokud ne, vrati se "None" typ. { Deprecated: Pokud ne,
## vrati se cela nafiltrovana data, ktera do funkce prisla
## (nedojde k vraceni specifickych objektu). }
if len(self.arrSeed) > 0:

## Zbaveni se duplikatu.
self.arrSeed = list( set ( self.arrSeed ) )
self.arrSeed = list(set(self.arrSeed))
print 'Hodnoty seedu: '
print self.arrSeed

Expand Down Expand Up @@ -175,7 +182,7 @@ def __init__(self, data, voxel, threshold = -1, interactivity = True,
## Zalozeni mist pro slidery
self.axcolor = 'white' # lightgoldenrodyellow
self.axmin = self.fig.add_axes([0.20, 0.24, 0.55, 0.03], axisbg = self.axcolor)
self.axmax = self.fig.add_axes([0.20, 0.20, 0.55, 0.03], axisbg = self.axcolor)
self.axmax = self.fig.add_axes([0.20, 0.20, 0.55, 0.03], axisbg = self.axcolor)
self.axclosing = self.fig.add_axes([0.20, 0.16, 0.55, 0.03], axisbg = self.axcolor)
self.axopening = self.fig.add_axes([0.20, 0.12, 0.55, 0.03], axisbg = self.axcolor)
self.axsigma = self.fig.add_axes([0.20, 0.08, 0.55, 0.03], axisbg = self.axcolor)
Expand All @@ -186,26 +193,22 @@ def __init__(self, data, voxel, threshold = -1, interactivity = True,
minBinaryOpening = 0
minSigma = 0.00

self.smin = Slider(self.axmin, 'Minimal threshold ' + str(self.min0),
self.smin = Slider(self.axmin, 'Minimal threshold ' + str(self.min0),
self.min0, self.max0, valinit = self.min0, dragging = False)
self.smax = Slider(self.axmax, 'Maximal threshold ' + str(self.min0),
self.smax = Slider(self.axmax, 'Maximal threshold ' + str(self.min0),
self.min0, self.max0, valinit = self.max0, dragging = False)

if(self.ICBinaryClosingIterations >= 1):
self.sclose = Slider(self.axclosing, 'Binary closing ' + str(minBinaryClosing),
minBinaryClosing, 100, valinit = self.ICBinaryClosingIterations, dragging = False)
self.sclose = Slider(self.axclosing, 'Binary closing', minBinaryClosing, 100, valinit = self.ICBinaryClosingIterations, dragging = False)
else:
self.sclose = Slider(self.axclosing, 'Binary closing ' + str(minBinaryClosing),
minBinaryClosing, 100, valinit = 0, dragging = False)
self.sclose = Slider(self.axclosing, 'Binary closing', minBinaryClosing, 100, valinit = 0, dragging = False)

if(self.ICBinaryOpeningIterations >= 1):
self.sopen = Slider(self.axopening, 'Binary opening ' + str(minBinaryOpening),
minBinaryOpening, 100, valinit = self.ICBinaryOpeningIterations, dragging = False)
self.sopen = Slider(self.axopening, 'Binary opening',minBinaryOpening, 100, valinit = self.ICBinaryOpeningIterations, dragging = False)
else:
self.sopen = Slider(self.axopening, 'Binary opening ' + str(minBinaryOpening),
minBinaryOpening, 100, valinit = 0, dragging = False)
self.sopen = Slider(self.axopening, 'Binary opening', minBinaryOpening, 100, valinit = 0, dragging = False)

self.ssigma = Slider(self.axsigma, 'Sigma ' + str(minSigma), 0.00, self.number, valinit = self.inputSigma, dragging = False)
self.ssigma = Slider(self.axsigma, 'Sigma' + str(minSigma), 0.00, self.number, valinit = self.inputSigma, dragging = False)

## Funkce slideru pri zmene jeho hodnoty
self.smin.on_changed(self.updateImage)
Expand Down Expand Up @@ -317,68 +320,39 @@ def updateImage(self, val):
sigmaNew = self.calculateSigma(sigma)

## Filtrovani
if(self.lastSigma != sigma):
scipy.ndimage.filters.gaussian_filter(self.imgFiltering, sigmaNew, 0, self.imgFiltering, 'reflect', 0.0)
scipy.ndimage.filters.gaussian_filter(self.imgFiltering, sigmaNew, order = 0, output = self.imgFiltering, mode = 'nearest')

if (self.lastSigma != sigma) or (self.threshold < 0):

self.calculateAutomaticThreshold()
## Ulozeni posledni hodnoty sigma pro neopakovani stejne operace
self.lastSigma = sigma
self.overrideSigma = True

else:

self.threshold = self.smin.val

del(sigmaNew)

## ====================================
## Prahovani (smin, smax)
## ====================================

if ( self.firstRun == True ) :

if ( self.threshold < 0 ) :

self.calculateAutomaticThreshold()

self.imgFiltering = self.imgFiltering * (self.imgFiltering >= self.threshold)
self.imgFiltering = self.imgFiltering * (self.imgFiltering <= self.smax.val)

if ( self.interactivity == True ) :

self.smin.val = (numpy.round(self.threshold, 2))
self.smin.valtext.set_text('{}'.format(self.smin.val))
self.smax.valtext.set_text('{}'.format(self.smax.val))
self.lastMinThresNum = self.smin.val
self.imgFiltering = self.imgFiltering * (self.imgFiltering >= self.threshold)
self.imgFiltering = self.imgFiltering * (self.imgFiltering <= self.smax.val)

self.overrideThres = True
if (self.interactivity == True) :

else:

if ( self.newThreshold == True ) :

if ( (self.lastMinThresNum != self.threshold) or (self.lastMaxThresNum != self.smax.val) or (self.overrideSigma == True) ):

self.imgFiltering = self.imgFiltering * (self.imgFiltering >= self.threshold)
self.imgFiltering = self.imgFiltering * (self.imgFiltering <= self.smax.val)

if ( self.interactivity == True ) :

self.smin.val = (numpy.round(self.threshold, 2))
self.smin.valtext.set_text('{}'.format(self.smin.val))

self.lastMinThresNum = self.threshold
self.lastMaxThresNum = self.smax.val
self.overrideThres = True

elif ( (self.lastMinThresNum != self.smin.val) or (self.lastMaxThresNum != self.smax.val) or (self.overrideSigma == True) ) :

self.imgFiltering = self.imgFiltering * (self.imgFiltering >= self.smin.val)
self.imgFiltering = self.imgFiltering * (self.imgFiltering <= self.smax.val)
self.lastMinThresNum = self.smin.val
self.lastMaxThresNum = self.smax.val
self.smin.val = (numpy.round(self.threshold, 2))
self.smin.valtext.set_text('{}'.format(self.smin.val))
self.smax.val = (numpy.round(self.smax.val, 2))
self.smax.valtext.set_text('{}'.format(self.smax.val))

## ====================================
## Operace binarni otevreni a uzavreni.
## ====================================

## Nastaveni hodnot slideru.
if ( self.interactivity == True ) :
if (self.interactivity == True) :

closeNum = int(numpy.round(self.sclose.val, 0))
openNum = int(numpy.round(self.sopen.val, 0))
Expand All @@ -391,21 +365,14 @@ def updateImage(self, val):
openNum = self.ICBinaryOpeningIterations

## Vlastni binarni uzavreni.
if(self.firstRun == True and self.ICBinaryClosingIterations >= 1) and closeNum > 0:
if (closeNum >= 1):

self.imgFiltering = self.imgFiltering * scipy.ndimage.binary_closing(self.imgFiltering, structure = None, iterations = closeNum)
else:
if( ((closeNum >= 1) and (self.lastCloseNum != closeNum)) or (self.overrideThres == True) ) and closeNum > 0:
self.imgFiltering = self.imgFiltering * scipy.ndimage.binary_closing(self.imgFiltering, structure = None, iterations = closeNum)
self.lastCloseNum = closeNum

## Vlastni binarni otevreni.
if(self.firstRun == True and self.ICBinaryOpeningIterations >= 1) and openNum > 0:
if (openNum >= 1):

self.imgFiltering = self.imgFiltering * scipy.ndimage.binary_opening(self.imgFiltering, structure = None, iterations = openNum)
else:
if( ((openNum >= 1) and (self.lastOpenNum != openNum)) or (self.overrideThres == True) ) and openNum > 0:
self.imgFiltering = self.imgFiltering * scipy.ndimage.binary_opening(self.imgFiltering,
structure = None, iterations = openNum)
self.lastOpenNum = openNum

## ====================================
## ====================================
Expand All @@ -414,9 +381,6 @@ def updateImage(self, val):
if (self.biggestObjects == True or self.seeds != None) :
self.imgFiltering = segmentation.getPriorityObjects(self.imgFiltering, self.nObj, self.seeds)

if self.imgFiltering != None:
print 'Typ dat k vykresleni: ' + str(type(self.imgFiltering[0][0][0]))

if (self.interactivity == True):

## Predani dat k vykresleni
Expand All @@ -428,6 +392,7 @@ def updateImage(self, val):

else:

print '(DEBUG) Typ dat k vykresleni: ' + str(type(self.imgFiltering[0][0][0]))
self.ax1.imshow(numpy.amax(self.imgFiltering, axis = 0, keepdims = self.numpyAMaxKeepDims), self.cmap)
self.ax2.imshow(numpy.amax(self.imgFiltering, axis = 1, keepdims = self.numpyAMaxKeepDims), self.cmap)
self.ax3.imshow(numpy.amax(self.imgFiltering, axis = 2, keepdims = self.numpyAMaxKeepDims), self.cmap)
Expand Down Expand Up @@ -456,7 +421,7 @@ def updateImage(self, val):

def calculateSigma(self, input):

if ( self.voxel[0] == self.voxel[1] == self.voxel[2] ):
if (self.voxel[0] == self.voxel[1] == self.voxel[2]):
return ((5 / self.voxel[0]) * input) / self.voxelV
else:
sigmaX = (5.0 / self.voxel[0]) * input
Expand Down Expand Up @@ -484,7 +449,8 @@ def calculateAutomaticThreshold(self):
pointsTo = 20 #(int)(hist_points * 0.1)
## Pocet bodu k preskoceni od konce hist
pointsSkip = (int)(hist_points * 0.025)
## hledani maxima: zacina se od 'start'*10 procent delky histu (aby se preskocili prvni oscilace)
## hledani maxima: zacina se od 'start'*10 procent delky histu (aby se
## preskocili prvni oscilace)
start = 0.1

## hist: funkce(threshold)
Expand Down Expand Up @@ -539,7 +505,8 @@ def calculateAutomaticThreshold(self):
x2 = []
y2 = []
place = 0
for index in range(init_index + pointsFrom + pointsSkip, init_index + pointsFrom + pointsSkip + pointsTo): # len(muj_histogram) - 5 - int(0.1 * len(muj_histogram)), len(muj_histogram) - 5
for index in range(init_index + pointsFrom + pointsSkip, init_index + pointsFrom + pointsSkip + pointsTo): # len(muj_histogram) - 5 - int(0.1 * len(muj_histogram)), len(muj_histogram) -
# 5
x2.insert(place, bin_centers[index])
y2.insert(place, muj_histogram[index])
## print("[ " + str(x2[place]) + ", " + str(y2[place]) + " ]")
Expand Down Expand Up @@ -708,16 +675,3 @@ def buttonPrevClosing(self, event):
self.fig.canvas.draw()
self.updateImage(0)














0 comments on commit 6354ac4

Please sign in to comment.