def _calculateMatrix(self):
P_gldm, angles = cMatrices.calculate_gldm(
self.matrix, self.maskArray,
numpy.array(self.settings.get('distances', [1])),
self.coefficients['Ng'], self.gldm_a,
self.settings.get('force2D', False),
self.settings.get('force2Ddimension', 0))
jvector = numpy.arange(1, P_gldm.shape[1] + 1, dtype='float64')
# Delete rows and columns that specify gray levels not present in the ROI
pd = numpy.sum(P_gldm, 0)
pg = numpy.sum(P_gldm, 1)
P_gldm = numpy.delete(P_gldm, numpy.where(pg == 0), 0)
P_gldm = numpy.delete(P_gldm, numpy.where(pd == 0), 1)
jvector = numpy.delete(jvector, numpy.where(pd == 0))
pg = numpy.delete(pg, numpy.where(pg == 0))
pd = numpy.delete(pd, numpy.where(pd == 0))
self.coefficients['pd'] = pd
self.coefficients['pg'] = pg
self.coefficients['ivector'] = self.coefficients['grayLevels']
self.coefficients['jvector'] = jvector
return P_gldm
def _calculateCMatrix(self):
angles = imageoperations.generateAngles(self.boundingBoxSize,
**self.kwargs)
Ng = self.coefficients['Ng']
P_gldm = cMatrices.calculate_gldm(self.matrix, self.maskArray, angles,
Ng, self.gldm_a)
jvector = numpy.arange(1, P_gldm.shape[1] + 1, dtype='float64')
# Delete rows and columns that specify gray levels not present in the ROI
sumP_gldm = numpy.sum(P_gldm)
pd = numpy.sum(P_gldm, 0)
pg = numpy.sum(P_gldm, 1)
P_gldm = numpy.delete(P_gldm, numpy.where(pg == 0), 0)
P_gldm = numpy.delete(P_gldm, numpy.where(pd == 0), 1)
jvector = numpy.delete(jvector, numpy.where(pd == 0))
pg = numpy.delete(pg, numpy.where(pg == 0))
pd = numpy.delete(pd, numpy.where(pd == 0))
self.coefficients['sumP_gldm'] = sumP_gldm
self.coefficients['pd'] = pd
self.coefficients['pg'] = pg
self.coefficients['ivector'] = self.coefficients['grayLevels']
self.coefficients['jvector'] = jvector
return P_gldm
def _calculateMatrix(self, voxelCoordinates=None):
self.logger.debug('Calculating GLDM matrix in C')
Ng = self.coefficients['Ng']
matrix_args = [
self.imageArray, self.maskArray,
numpy.array(self.settings.get('distances', [1])), Ng, self.gldm_a,
self.settings.get('force2D', False),
self.settings.get('force2Ddimension', 0)
]
if self.voxelBased:
matrix_args += [
self.settings.get('kernelRadius', 1), voxelCoordinates
]
P_gldm = cMatrices.calculate_gldm(*matrix_args) # shape (Nv, Ng, Nd)
# Delete rows that specify gray levels not present in the ROI
NgVector = range(1, Ng + 1) # All possible gray values
GrayLevels = self.coefficients[
'grayLevels'] # Gray values present in ROI
emptyGrayLevels = numpy.array(
list(set(NgVector) - set(GrayLevels)),
dtype=int) # Gray values NOT present in ROI
if list(emptyGrayLevels):
P_gldm = numpy.delete(P_gldm, emptyGrayLevels - 1, 1)
jvector = numpy.arange(1, P_gldm.shape[2] + 1, dtype='float64')
# shape (Nv, Nd)
pd = numpy.sum(P_gldm, 1)
# shape (Nv, Ng)
pg = numpy.sum(P_gldm, 2)
# Delete columns that dependence sizes not present in the ROI
empty_sizes = numpy.sum(pd, 0)
P_gldm = numpy.delete(P_gldm, numpy.where(empty_sizes == 0), 2)
jvector = numpy.delete(jvector, numpy.where(empty_sizes == 0))
pd = numpy.delete(pd, numpy.where(empty_sizes == 0), 1)
Nz = numpy.sum(pd, 1) # Nz per kernel, shape (Nv, )
Nz[Nz == 0] = 1 # set sum to numpy.spacing(1) if sum is 0?
self.coefficients['Nz'] = Nz
self.coefficients['pd'] = pd
self.coefficients['pg'] = pg
self.coefficients['ivector'] = self.coefficients['grayLevels'].astype(
float)
self.coefficients['jvector'] = jvector
return P_gldm
def _calculateMatrix(self):
self.logger.debug('Calculating GLDM matrix in C')
Ng = self.coefficients['Ng']
P_gldm = cMatrices.calculate_gldm(
self.matrix, self.maskArray,
numpy.array(self.settings.get('distances', [1])), Ng, self.gldm_a,
self.settings.get('force2D', False),
self.settings.get('force2Ddimension', 0))
# Delete rows that specify gray levels not present in the ROI
NgVector = range(1, Ng + 1) # All possible gray values
GrayLevels = self.coefficients[
'grayLevels'] # Gray values present in ROI
emptyGrayLevels = numpy.array(
list(set(NgVector) -
set(GrayLevels))) # Gray values NOT present in ROI
P_gldm = numpy.delete(P_gldm, emptyGrayLevels - 1, 0)
jvector = numpy.arange(1, P_gldm.shape[1] + 1, dtype='float64')
pd = numpy.sum(P_gldm, 0)
pg = numpy.sum(P_gldm, 1)
# Delete columns that dependence sizes not present in the ROI
P_gldm = numpy.delete(P_gldm, numpy.where(pd == 0), 1)
jvector = numpy.delete(jvector, numpy.where(pd == 0))
pd = numpy.delete(pd, numpy.where(pd == 0))
self.coefficients['pd'] = pd
self.coefficients['pg'] = pg
self.coefficients['ivector'] = self.coefficients['grayLevels']
self.coefficients['jvector'] = jvector
return P_gldm