# print "Reading pre-built classifier from file::" , classifierFilename , "@" , datetime.datetime.now()
# classifier = loadClassifier(classifierFilename)
#
# predictImage = pomio.msrc_loadImages(msrcData)[0]
plt.interactive(1)
plt.figure()
pomio.showClassColours()
plt.figure()
for img in msrcImages:
print "\nRead in image %s from the MSRC dataset::" % img.m_imgFn
imageFeatures = FeatureGenerator.generatePixelFeaturesForImage(img.m_img)
predLabs = classifier.predict(imageFeatures)
print "\nGenerating prediction of shape ", predLabs.shape, "::" , predLabs
predImg = np.reshape( predLabs, img.m_img[:,:,0].shape )
#print "\nGenerating the probability dist. for each pixel over class labels @" , datetime.datetime.now()
#imageClassDist = generateImagePredictionClassDist(img.m_img, classifier,False)
# Display
print "Unique labels from clfr = ", np.unique(predLabs)
plt.subplot(1,2,1)
plt.imshow(img.m_img)
plt.subplot(1,2,2)
pomio.showLabels( predImg )
plt.waitforbuttonpress()
print "\tCompleted @ " + str(datetime.datetime.now())
plt.interactive(0)
plt.show()
plt.figure()
pomio.showClassColours()
plt.figure()
print 'Classifying file ', args.infile
image = skimage.io.imread(args.infile)
[spClassPreds, spGraph
] = SuperPixelClassifier.predictSuperPixelLabels(clfr, image,
numberSuperPixels,
superPixelCompactness)
spClassPredsImage = spGraph.imageFromSuperPixelData(
spClassPreds.reshape((len(spClassPreds), 1)))
if args.verbose:
plt.subplot(1, 2, 1)
plt.imshow(image)
plt.title(args.infile)
plt.subplot(1, 2, 2)
pomio.showLabels(spClassPredsImage)
plt.waitforbuttonpress()
if args.outfile and len(args.outfile) > 0:
print 'Writing output label file %s' % args.outfile
outimg = pomio.msrc_convertLabelsToRGB(spClassPredsImage)
skimage.io.imsave(args.outfile, outimg)
print ' done.'
if args.verbose:
plt.interactive(0)
plt.show()
plt.figure()
plt.imshow(imgRGB)
plt.title('original image')
# show superpixels
plt.figure()
spix.draw()
plt.title('Super Pixels')
# show class labels per pixel
classLabs = np.argmax( classProbs, 1 )
# these are labs per region. turn into an image of labels.
plt.figure()
# +1 adds void class
# todo: tidy this shemozzle up!
pomio.showLabels( spix.imageFromSuperPixelData( classLabs + 1 ) )
plt.title('Raw Classifier Labelling')
plt.figure()
pomio.showClassColours()
plt.draw()
if dointeract:
print 'Click plot to continue...'
plt.waitforbuttonpress()
#
# Do the inference
#
print 'Performing CRF inference...'
#
# predictImage = pomio.msrc_loadImages(msrcData)[0]
plt.interactive(1)
plt.figure()
pomio.showClassColours()
plt.figure()
for img in msrcImages:
print "\nRead in image %s from the MSRC dataset::" % img.m_imgFn
# todo: replace with features.computePixelFeatures JRS
imageFeatures = FeatureGenerator.generatePixelFeaturesForImage(
img.m_img)
predLabs = classifier.predict(imageFeatures)
print "\nGenerating prediction of shape ", predLabs.shape, "::", predLabs
predImg = np.reshape(predLabs, img.m_img[:, :, 0].shape)
#print "\nGenerating the probability dist. for each pixel over class labels @" , datetime.datetime.now()
#imageClassDist = generateImagePredictionClassDist(img.m_img, classifier,False)
# Display
print "Unique labels from clfr = ", np.unique(predLabs)
plt.subplot(1, 2, 1)
plt.imshow(img.m_img)
plt.subplot(1, 2, 2)
pomio.showLabels(predImg)
plt.waitforbuttonpress()
print "\tCompleted @ " + str(datetime.datetime.now())
plt.interactive(0)
plt.show()
print ' classes = ', clfr.classes_
# Transform class probs to the correct sized matrix.
nbRows = imgRGB.shape[0]
nbCols = imgRGB.shape[1]
nbClasses = pomio.getNumClasses()
cpnew = np.zeros( (nbRows, nbCols, nbClasses) )
for i in range( classProbs.shape[2] ):
# stuff this set of probs to new label
cpnew[:,:,clfr.classes_[i]] = classProbs[:,:,i]
classProbs = cpnew
del cpnew
maxLabel = np.argmax( classProbs, 2 )
pomio.showLabels(maxLabel, colourMap)
if args.verbose:
plt.title('raw clfr labels')
plt.figure()
pomio.showClassColours( classNames, colourMap )
plt.draw()
if 0 and args.interactive:
plt.waitforbuttonpress()
#print classProbs
if dbgMode and args.verbose:
for i in range( classProbs.shape[2] ):
plt.imshow( classProbs[:,:,i] )
plt.title( 'class %d: %s' % (i,classNames[i]) )
print ' classes = ', clfr.classes_
# Transform class probs to the correct sized matrix.
nbRows = imgRGB.shape[0]
nbCols = imgRGB.shape[1]
nbClasses = pomio.getNumClasses()
cpnew = np.zeros((nbRows, nbCols, nbClasses))
for i in range(classProbs.shape[2]):
# stuff this set of probs to new label
cpnew[:, :, clfr.classes_[i]] = classProbs[:, :, i]
classProbs = cpnew
del cpnew
maxLabel = np.argmax(classProbs, 2)
pomio.showLabels(maxLabel)
plt.title('raw clfr labels')
plt.figure()
pomio.showClassColours()
plt.draw()
if dointeract:
plt.waitforbuttonpress()
print classProbs
if dbgMode:
for i in range(classProbs.shape[2]):
plt.imshow(classProbs[:, :, i])
plt.title('class %d: %s' % (i, pomio.msrc_classLabels[i]))
plt.waitforbuttonpress()
plt.figure()
plt.imshow(imgRGB)
plt.title('original image')
# show superpixels
plt.figure()
spix.draw()
plt.title('Super Pixels')
# show class labels per pixel
classLabs = np.argmax(classProbs, 1)
# these are labs per region. turn into an image of labels.
plt.figure()
pomio.showLabels(spix.imageFromSuperPixelData(classLabs))
plt.title('Raw Classifier Labelling')
plt.figure()
pomio.showClassColours()
plt.draw()
if dointeract:
print 'Click plot to continue...'
plt.waitforbuttonpress()
#
# Do the inference
#
# Get adjacency probs
if args.adjFn != None:
plt.figure()
plt.imshow(imgRGB)
plt.title('original image')
# show superpixels
plt.figure()
spix.draw()
plt.title('Super Pixels')
# show class labels per pixel
classLabs = np.argmax(classProbs, 1)
# these are labs per region. turn into an image of labels.
plt.figure()
# +1 adds void class
# todo: tidy this shemozzle up!
pomio.showLabels(spix.imageFromSuperPixelData(classLabs + 1))
plt.title('Raw Classifier Labelling')
plt.figure()
pomio.showClassColours()
plt.draw()
if dointeract:
print 'Click plot to continue...'
plt.waitforbuttonpress()
#
# Do the inference
#
print 'Performing CRF inference...'
if args.verbose:
plt.waitforbuttonpress()
plt.figure()
plt.imshow(imgRGB)
plt.title('original image')
# show superpixels
plt.figure()
spix.draw()
plt.title('Super Pixels')
# show class labels per pixel
classLabs = np.argmax( classProbs, 1 )
# these are labs per region. turn into an image of labels.
plt.figure()
pomio.showLabels( spix.imageFromSuperPixelData( classLabs ) )
plt.title('Raw Classifier Labelling')
plt.figure()
pomio.showClassColours()
plt.draw()
if dointeract:
print 'Click plot to continue...'
plt.waitforbuttonpress()
#
# Do the inference
#
# Get adjacency probs
if args.adjFn != None:
# Load all data
print 'Loading all data...'
# first arg should be msrc data path
data = pomio.msrc_loadImages( sys.argv[1], ['Images/7_3_s.bmp'])
# get particular image we like
ex = data[0]
plt.figure()
plt.imshow(ex.m_img)
plt.title('original image')
plt.figure()
clrs = [[z/255.0 for z in c[1]] for c in pomio.msrc_classToRGB]
pomio.showLabels( ex.m_gt )
plt.title('ground truth labels' )
print 'unique class labels: ', np.unique(ex.m_gt)
# generate features
imagePixelFeatures = FeatureGenerator.generatePixelFeaturesForImage(ex.m_img)
# For each feature, how many distinct values?
for i in range(imagePixelFeatures.shape[1]):
print "Feature %d has %d distinct values" \
% (i, len(np.unique( imagePixelFeatures[:,i])) )
# Plot a selection of features
sel = np.arange(26,30)
# sel = range(80,86)
superPixelCompactness = args.superPixelCompactness
if args.verbose:
plt.interactive(1)
plt.figure()
pomio.showClassColours()
plt.figure()
print 'Classifying file ', args.infile
image = skimage.io.imread(args.infile)
[spClassPreds, spGraph] = SuperPixelClassifier.predictSuperPixelLabels(clfr, image,numberSuperPixels, superPixelCompactness)
spClassPredsImage = spGraph.imageFromSuperPixelData( spClassPreds.reshape( (len(spClassPreds),1) ) )
if args.verbose:
plt.subplot(1,2,1)
plt.imshow(image)
plt.title(args.infile)
plt.subplot(1,2,2)
pomio.showLabels(spClassPredsImage)
plt.waitforbuttonpress()
if args.outfile and len(args.outfile)>0:
print 'Writing output label file %s' % args.outfile
outimg = pomio.msrc_convertLabelsToRGB( spClassPredsImage )
skimage.io.imsave(args.outfile, outimg)
print ' done.'
if args.verbose:
plt.interactive(0)
plt.show()
print ' classes = ', clfr.classes_
# Transform class probs to the correct sized matrix.
nbRows = imgRGB.shape[0]
nbCols = imgRGB.shape[1]
nbClasses = pomio.getNumClasses()
cpnew = np.zeros( (nbRows, nbCols, nbClasses) )
for i in range( classProbs.shape[2] ):
# stuff this set of probs to new label
cpnew[:,:,clfr.classes_[i]] = classProbs[:,:,i]
classProbs = cpnew
del cpnew
maxLabel = np.argmax( classProbs, 2 )
pomio.showLabels(maxLabel)
plt.title('raw clfr labels')
plt.figure()
pomio.showClassColours()
plt.draw()
if dointeract:
plt.waitforbuttonpress()
print classProbs
if dbgMode:
for i in range( classProbs.shape[2] ):
plt.imshow( classProbs[:,:,i] )
plt.title( 'class %d: %s' % (i,pomio.msrc_classLabels[i]) )
import matplotlib
# Load all data
print 'Loading all data...'
data = pomio.msrc_loadImages('/home/jamie/data/MSRC_ObjCategImageDatabase_v2', \
['Images/7_3_s.bmp'])
# get particular image we like
ex = data[0]
plt.figure()
plt.imshow(ex.m_img)
plt.title('original image')
plt.figure()
clrs = [[z / 255.0 for z in c[1]] for c in pomio.msrc_classToRGB]
pomio.showLabels(ex.m_gt)
plt.title('ground truth labels')
print 'unique class labels: ', np.unique(ex.m_gt)
# generate features
imagePixelFeatures = FeatureGenerator.generatePixelFeaturesForImage(ex.m_img)
# For each feature, how many distinct values?
for i in range(imagePixelFeatures.shape[1]):
print "Feature %d has %d distinct values" \
% (i, len(np.unique( imagePixelFeatures[:,i])) )
# Plot a selection of features
sel = np.arange(12, 29, 2)
# sel = range(80,86)
