def precond_proximal(self, x, tau, out=None):
"""Modify proximal method to work with preconditioned tau"""
pars = {
'algorithm':
FGP_TV,
'input':
np.asarray(x.as_array() / tau.as_array(), dtype=np.float32),
'regularization_parameter':
self.lambdaReg,
'number_of_iterations':
self.iterationsTV,
'tolerance_constant':
self.tolerance,
'methodTV':
self.methodTV,
'nonneg':
self.nonnegativity,
'printingOut':
self.printing
}
res, info = regularisers.FGP_TV(pars['input'],
pars['regularization_parameter'],
pars['number_of_iterations'],
pars['tolerance_constant'],
pars['methodTV'], pars['nonneg'],
self.device)
if out is not None:
out.fill(res)
else:
out = x.copy()
out.fill(res)
out *= tau
return out
def proximal_numpy(self, in_arr, tau, out = None):
res , info = regularisers.FGP_TV(\
in_arr,\
self.alpha*tau,\
self.max_iteration,\
self.tolerance,\
self.methodTV,\
self.nonnegativity,\
self.device)
return res, info
def test_functionality_FGP_TV(self):
data = dataexample.CAMERA.get(size=(256,256))
datarr = data.as_array()
from cil.plugins.ccpi_regularisation.functions import FGP_TV
from ccpi.filters import regularisers
tau = 1.
fcil = FGP_TV()
outcil = fcil.proximal(data, tau=tau)
# use CIL defaults
outrgl, info = regularisers.FGP_TV(datarr, fcil.alpha*tau, fcil.max_iteration, fcil.tolerance, 0, 1, 'cpu' )
np.testing.assert_almost_equal(outrgl, outcil.as_array())
def proximal(self, x, tau, out=None):
in_arr = np.asarray(x.as_array(), dtype=np.float32, order='C')
res , info = regularisers.FGP_TV(\
in_arr,\
self.alpha*tau,\
self.max_iteration,\
self.tolerance,\
self.methodTV,\
self.nonnegativity,\
self.device)
if out is not None:
out.fill(res)
else:
out = x.copy()
out.fill(res)
return out
def prox(self,x,tau):
pars = {'algorithm' : FGP_TV, \
'input' : np.asarray(x.as_array(), dtype=np.float32),\
'regularization_parameter':self.lambdaReg*tau, \
'number_of_iterations' :self.iterationsTV ,\
'tolerance_constant':self.tolerance,\
'methodTV': self.methodTV ,\
'nonneg': self.nonnegativity ,\
'printingOut': self.printing}
out = regularisers.FGP_TV(pars['input'],
pars['regularization_parameter'],
pars['number_of_iterations'],
pars['tolerance_constant'],
pars['methodTV'],
pars['nonneg'],
pars['printingOut'], self.device)
return DataContainer(out)
def proximal(self, x, tau, out=None):
pars = {'algorithm' : FGP_TV, \
'input' : np.asarray(x.as_array(), dtype=np.float32),\
'regularization_parameter':self.lambdaReg*tau, \
'number_of_iterations' :self.iterationsTV ,\
'tolerance_constant':self.tolerance,\
'methodTV': self.methodTV ,\
'nonneg': self.nonnegativity ,\
'printingOut': self.printing}
res, info = regularisers.FGP_TV(pars['input'],
pars['regularization_parameter'],
pars['number_of_iterations'],
pars['tolerance_constant'],
pars['methodTV'], pars['nonneg'],
self.device)
if out is not None:
out.fill(res)
else:
out = x.copy()
out.fill(res)
return out