def par_xstep(i):
r"""Minimise Augmented Lagrangian with respect to
:math:`\mathbf{x}_{G_i}`, one of the disjoint problems of optimizing
:math:`\mathbf{x}`.
Parameters
----------
i : int
Index of grouping to update
"""
global mp_X
global mp_DX
YU0f = sl.rfftn(mp_Y0[[i]] - mp_U0[[i]], mp_Nv, mp_axisN)
YU1f = sl.rfftn(mp_Y1[mp_grp[i]:mp_grp[i+1]] -
1/mp_alpha*mp_U1[mp_grp[i]:mp_grp[i+1]], mp_Nv, mp_axisN)
if mp_Cd == 1:
b = np.conj(mp_Df[mp_grp[i]:mp_grp[i+1]]) * YU0f + mp_alpha**2*YU1f
Xf = sl.solvedbi_sm(mp_Df[mp_grp[i]:mp_grp[i+1]], mp_alpha**2, b,
mp_cache[i], axis=mp_axisM)
else:
b = sl.inner(np.conj(mp_Df[mp_grp[i]:mp_grp[i+1]]), YU0f,
axis=mp_C) + mp_alpha**2*YU1f
Xf = sl.solvemdbi_ism(mp_Df[mp_grp[i]:mp_grp[i+1]], mp_alpha**2, b,
mp_axisM, mp_axisC)
mp_X[mp_grp[i]:mp_grp[i+1]] = sl.irfftn(Xf, mp_Nv,
mp_axisN)
mp_DX[i] = sl.irfftn(sl.inner(mp_Df[mp_grp[i]:mp_grp[i+1]], Xf,
mp_axisM), mp_Nv, mp_axisN)
def ccmodmd_xstep(k):
"""Do the X step of the ccmod stage. The only parameter is the slice
index `k` and there are no return values; all inputs and outputs are
from and to global variables.
"""
YU0 = mp_D_Y0 - mp_D_U0[k]
YU1 = mp_D_Y1[k] + mp_S[k] - mp_D_U1[k]
b = sl.rfftn(YU0, None, mp_cri.axisN) + \
np.conj(mp_Zf[k]) * sl.rfftn(YU1, None, mp_cri.axisN)
Xf = sl.solvedbi_sm(mp_Zf[k], 1.0, b, axis=mp_cri.axisM)
mp_D_X[k] = sl.irfftn(Xf, mp_cri.Nv, mp_cri.axisN)
mp_DX[k] = sl.irfftn(sl.inner(Xf, mp_Zf[k]), mp_cri.Nv, mp_cri.axisN)
def xstep(self):
r"""Minimise Augmented Lagrangian with respect to
:math:`\mathbf{x}`.
"""
self.YU[:] = self.Y - self.U
self.block_sep0(self.YU)[:] += self.S
Zf = sl.rfftn(self.YU, None, self.cri.axisN)
Z0f = self.block_sep0(Zf)
Z1f = self.block_sep1(Zf)
DZ0f = np.conj(self.Df) * Z0f
DZ0fBQ = sl.dot(self.B.dot(self.Q).T, DZ0f, axis=self.cri.axisC)
Z1fQ = sl.dot(self.Q.T, Z1f, axis=self.cri.axisC)
b = DZ0fBQ + Z1fQ
Xh = sl.solvedbd_sm(self.gDf, (self.mu / self.rho) * self.GHGf + 1.0,
b, self.c, axis=self.cri.axisM)
self.Xf[:] = sl.dot(self.Q, Xh, axis=self.cri.axisC)
self.X = sl.irfftn(self.Xf, self.cri.Nv, self.cri.axisN)
if self.opt['LinSolveCheck']:
DDXf = np.conj(self.Df) * sl.inner(self.Df, self.Xf,
axis=self.cri.axisM)
DDXfBB = sl.dot(self.B.T.dot(self.B), DDXf, axis=self.cri.axisC)
ax = self.rho * (DDXfBB + self.Xf) + \
self.mu * self.GHGf * self.Xf
b = self.rho * (sl.dot(self.B.T, DZ0f, axis=self.cri.axisC)
+ Z1f)
self.xrrs = sl.rrs(ax, b)
else:
self.xrrs = None
def xstep(self):
r"""Minimise Augmented Lagrangian with respect to
:math:`\mathbf{x}`."""
self.YU[:] = self.Y - self.U
YUf = sl.rfftn(self.YU, None, self.cri.axisN)
# The sum is over the extra axis indexing spatial gradient
# operators G_i, *not* over axisM
b = self.DSf + self.rho*(YUf[..., -1] + self.Wtv * np.sum(
np.conj(self.Gf) * YUf[..., 0:-1], axis=-1))
if self.cri.Cd == 1:
self.Xf[:] = sl.solvedbi_sm(
self.Df, self.rho*self.GHGf + self.rho, b, self.c,
self.cri.axisM)
else:
self.Xf[:] = sl.solvemdbi_ism(
self.Df, self.rho*self.GHGf + self.rho, b, self.cri.axisM,
self.cri.axisC)
self.X = sl.irfftn(self.Xf, self.cri.Nv, self.cri.axisN)
if self.opt['LinSolveCheck']:
Dop = lambda x: sl.inner(self.Df, x, axis=self.cri.axisM)
if self.cri.Cd == 1:
DHop = lambda x: np.conj(self.Df) * x
else:
DHop = lambda x: sl.inner(np.conj(self.Df), x,
axis=self.cri.axisC)
ax = DHop(Dop(self.Xf)) + (self.rho*self.GHGf + self.rho)*self.Xf
self.xrrs = sl.rrs(ax, b)
else:
self.xrrs = None
def xstep(self):
r"""Minimise Augmented Lagrangian with respect to block vector
:math:`\mathbf{x} = \left( \begin{array}{ccc} \mathbf{x}_0^T &
\mathbf{x}_1^T & \ldots \end{array} \right)^T\;`.
"""
# This test reflects empirical evidence that two slightly
# different implementations are faster for single or
# multi-channel data. This kludge is intended to be temporary.
if self.cri.Cd > 1:
for i in range(self.Nb):
self.xistep(i)
else:
self.YU[:] = self.Y[..., np.newaxis] - self.U
b = np.swapaxes(self.ZSf[..., np.newaxis], self.cri.axisK, -1) \
+ self.rho*sl.rfftn(self.YU, None, self.cri.axisN)
for i in range(self.Nb):
self.Xf[..., i] = sl.solvedbi_sm(
self.Zf[..., [i], :], self.rho, b[..., i],
axis=self.cri.axisM)
self.X = sl.irfftn(self.Xf, self.cri.Nv, self.cri.axisN)
if self.opt['LinSolveCheck']:
ZSfs = np.sum(self.ZSf, axis=self.cri.axisK, keepdims=True)
YU = np.sum(self.Y[..., np.newaxis] - self.U, axis=-1)
b = ZSfs + self.rho*sl.rfftn(YU, None, self.cri.axisN)
Xf = self.swapaxes(self.Xf)
Zop = lambda x: sl.inner(self.Zf, x, axis=self.cri.axisM)
ZHop = lambda x: np.conj(self.Zf) * x
ax = np.sum(ZHop(Zop(Xf)) + self.rho*Xf, axis=self.cri.axisK,
keepdims=True)
self.xrrs = sl.rrs(ax, b)
else:
self.xrrs = None
def reconstruct(self, X=None):
"""Reconstruct representation."""
if X is None:
X = self.X
Xf = sl.rfftn(X, None, self.cri.axisN)
Sf = np.sum(self.Df * Xf, axis=self.cri.axisM)
return sl.irfftn(Sf, self.cri.Nv, self.cri.axisN)
def obfn_dfd(self):
r"""Compute data fidelity term :math:`(1/2) \| W \left( \sum_m
\mathbf{d}_m * \mathbf{x}_m - \mathbf{s} \right) \|_2^2`.
"""
Ef = sl.inner(self.Zf, self.obfn_fvarf(), axis=self.cri.axisM) \
- self.Sf
return (np.linalg.norm(self.W * sl.irfftn(Ef, self.cri.Nv,
self.cri.axisN))**2) / 2.0
def cnst_AT(self, X):
r"""Compute :math:`A^T \mathbf{x}` where :math:`A \mathbf{x}` is
a component of ADMM problem constraint. In this case
:math:`A^T \mathbf{x} = (G_r^T \;\; G_c^T) \mathbf{x}`.
"""
Xf = sl.rfftn(X, axes=self.axes)
return np.sum(sl.irfftn(np.conj(self.Gf)*Xf, self.axsz,
axes=self.axes), axis=self.Y.ndim-1)
def obfn_dfd(self):
r"""Compute data fidelity term :math:`(1/2) \| W \left( \sum_m
\mathbf{d}_m * \mathbf{x}_m - \mathbf{s} \right) \|_2^2`.
"""
XF = sl.rfftn(self.obfn_fvar(), mp_Nv, mp_axisN)
DX = np.moveaxis(sl.irfftn(sl.inner(mp_Df, XF, mp_axisM),
mp_Nv, mp_axisN), mp_axisM,
self.cri.axisM)
return np.sum((self.W*(DX-self.S))**2)/2.0
def cnst_A1T(self, Y1):
r"""Compute :math:`A_1^T \mathbf{y}_1` component of
:math:`A^T \mathbf{y}`. In this case :math:`A_1^T \mathbf{y}_1 =
(\Gamma_0^T \;\; \Gamma_1^T \;\; \ldots) \mathbf{y}_1`.
"""
Y1f = sl.rfftn(Y1, None, axes=self.cri.axisN)
return sl.irfftn(np.conj(self.GDf) * Y1f, self.cri.Nv,
self.cri.axisN)
def obfn_dfd(self):
r"""Compute data fidelity term :math:`(1/2) \sum_k \| W (\sum_m
\mathbf{d}_m * \mathbf{x}_{k,m} - \mathbf{s}_k) \|_2^2`
"""
Ef = self.eval_Rf(self.Xf)
E = sl.irfftn(Ef, self.cri.Nv, self.cri.axisN)
return (np.linalg.norm(self.W * E)**2) / 2.0
def cbpdnmd_xstep(k):
"""Do the X step of the cbpdn stage. The only parameter is the slice
index `k` and there are no return values; all inputs and outputs are
from and to global variables.
"""
YU0 = mp_Z_Y0[k] + mp_S[k] - mp_Z_U0[k]
YU1 = mp_Z_Y1[k] - mp_Z_U1[k]
if mp_cri.Cd == 1:
b = np.conj(mp_Df) * sl.rfftn(YU0, None, mp_cri.axisN) + \
sl.rfftn(YU1, None, mp_cri.axisN)
Xf = sl.solvedbi_sm(mp_Df, 1.0, b, axis=mp_cri.axisM)
else:
b = sl.inner(np.conj(mp_Df), sl.rfftn(YU0, None, mp_cri.axisN),
axis=mp_cri.axisC) + \
sl.rfftn(YU1, None, mp_cri.axisN)
Xf = sl.solvemdbi_ism(mp_Df, 1.0, b, mp_cri.axisM, mp_cri.axisC)
mp_Z_X[k] = sl.irfftn(Xf, mp_cri.Nv, mp_cri.axisN)
mp_DX[k] = sl.irfftn(sl.inner(mp_Df, Xf), mp_cri.Nv, mp_cri.axisN)
def cnst_A(self, X, Xf=None):
r"""Compute :math:`A \mathbf{x}` component of ADMM problem
constraint. In this case :math:`A \mathbf{x} = (G_r^T \;\;
G_c^T)^T \mathbf{x}`.
"""
if Xf is None:
Xf = sl.rfftn(X, axes=self.axes)
return sl.irfftn(self.Gf*Xf[..., np.newaxis], self.axsz,
axes=self.axes)
def cnst_A0T(self, X):
r"""Compute :math:`A_0^T \mathbf{x}` where :math:`A_0 \mathbf{x}`
is a component of the ADMM problem constraint. In this case
:math:`A_0^T \mathbf{x} = (\Gamma_0^T \;\; \Gamma_1^T \;\; \ldots )
\mathbf{x}`.
"""
Xf = sl.rfftn(X, axes=self.cri.axisN)
return self.Wtv[..., np.newaxis] * sl.irfftn(
np.conj(self.Gf) * Xf[..., 0:-1], self.cri.Nv, axes=self.cri.axisN)
def cnst_A0(self, X, Xf=None):
r"""Compute :math:`A_0 \mathbf{x}` component of ADMM problem
constraint. In this case :math:`A_0 \mathbf{x} = (\Gamma_0^T \;\;
\Gamma_1^T \;\; \ldots )^T \mathbf{x}`.
"""
if Xf is None:
Xf = sl.rfftn(X, axes=self.cri.axisN)
return self.Wtv[..., np.newaxis] * sl.irfftn(
self.Gf * Xf[..., np.newaxis], self.cri.Nv, axes=self.cri.axisN)
def cnst_A0(self, X, Xf=None):
r"""Compute :math:`A_0 \mathbf{x}` component of ADMM problem
constraint.
"""
if Xf is None:
Xf = sl.rfftn(X, None, self.cri.axisN)
return sl.irfftn(
sl.dot(self.B, sl.inner(self.Df, Xf, axis=self.cri.axisM),
axis=self.cri.axisC), self.cri.Nv, self.cri.axisN)
def ccmod_xstep(k):
"""Do the X step of the ccmod stage. The only parameter is the slice
index `k` and there are no return values; all inputs and outputs are
from and to global variables.
"""
YU = mp_D_Y - mp_D_U[k]
b = mp_ZSf[k] + mp_drho * sl.rfftn(YU, None, mp_cri.axisN)
Xf = sl.solvedbi_sm(mp_Zf[k], mp_drho, b, axis=mp_cri.axisM)
mp_D_X[k] = sl.irfftn(Xf, mp_cri.Nv, mp_cri.axisN)
def reconstruct(self, D=None):
"""Reconstruct representation."""
if D is None:
Df = self.Xf
else:
Df = sl.rfftn(D, None, self.cri.axisN)
Sf = np.sum(self.Zf * Df, axis=self.cri.axisM)
return sl.irfftn(Sf, self.cri.Nv, self.cri.axisN)
def xstep(self):
r"""Minimise Augmented Lagrangian with respect to :math:`\mathbf{x}`.
"""
self.YU[:] = self.Y - self.U
b = self.ZSf + self.rho*sl.rfftn(self.YU, None, self.cri.axisN)
self.Xf[:] = sl.solvemdbi_ism(self.Zf, self.rho, b, self.cri.axisM,
self.cri.axisK)
self.X = sl.irfftn(self.Xf, self.cri.Nv, self.cri.axisN)
self.xstep_check(b)
def compute_residuals(self):
"""Compute residuals and stopping thresholds. The parent class
method is overridden to ensure that the residual calculations
include the additional variables introduced in the modification
to the baseline algorithm.
"""
# The full primary residual is straightforward to compute from
# the primary residuals for the baseline algorithm and for the
# additional variables
r0 = self.rsdl_r(self.AXnr, self.Y)
r1 = self.AX1nr - self.Y1 - self.S
r = np.sqrt(np.sum(r0**2) + np.sum(r1**2))
# The full dual residual is more complicated to compute than the
# full primary residual
ATU = self.swapaxes(self.U) + sl.irfftn(
np.conj(self.Zf) * sl.rfftn(self.U1, self.cri.Nv, self.cri.axisN),
self.cri.Nv, self.cri.axisN)
s = self.rho * np.linalg.norm(ATU)
# The normalisation factor for the full primal residual is also not
# straightforward
nAX = np.sqrt(np.linalg.norm(self.AXnr)**2 +
np.linalg.norm(self.AX1nr)**2)
nY = np.sqrt(np.linalg.norm(self.Y)**2 +
np.linalg.norm(self.Y1)**2)
rn = max(nAX, nY, np.linalg.norm(self.S))
# The normalisation factor for the full dual residual is
# straightforward to compute
sn = self.rho * np.sqrt(np.linalg.norm(self.U)**2 +
np.linalg.norm(self.U1)**2)
# Final residual values and stopping tolerances depend on
# whether standard or normalised residuals are specified via the
# options object
if self.opt['AutoRho', 'StdResiduals']:
epri = np.sqrt(self.Nc)*self.opt['AbsStopTol'] + \
rn*self.opt['RelStopTol']
edua = np.sqrt(self.Nx)*self.opt['AbsStopTol'] + \
sn*self.opt['RelStopTol']
else:
if rn == 0.0:
rn = 1.0
if sn == 0.0:
sn = 1.0
r /= rn
s /= sn
epri = np.sqrt(self.Nc)*self.opt['AbsStopTol']/rn + \
self.opt['RelStopTol']
edua = np.sqrt(self.Nx)*self.opt['AbsStopTol']/sn + \
self.opt['RelStopTol']
return r, s, epri, edua
def reconstruct(self, D=None, X=None):
"""Reconstruct representation."""
if D is None:
D = self.getdict(crop=False)
if X is None:
X = self.getcoef()
Df = sl.rfftn(D, self.xstep.cri.Nv, self.xstep.cri.axisN)
Xf = sl.rfftn(X, self.xstep.cri.Nv, self.xstep.cri.axisN)
DXf = sl.inner(Df, Xf, axis=self.xstep.cri.axisM)
return sl.irfftn(DXf, self.xstep.cri.Nv, self.xstep.cri.axisN)
def cnst_A0T(self, Y0):
r"""Compute :math:`A_0^T \mathbf{y}_0` component of
:math:`A^T \mathbf{y}` (see :meth:`.ADMMTwoBlockCnstrnt.cnst_AT`).
"""
# This calculation involves non-negligible computational cost. It
# should be possible to disable relevant diagnostic information
# (dual residual) to avoid this cost.
Y0f = sl.rfftn(Y0, None, self.cri.axisN)
return sl.irfftn(
sl.dot(self.B.T, np.conj(self.Df) * Y0f, axis=self.cri.axisC),
self.cri.Nv, self.cri.axisN)
def cnst_A0(self, X, Xf=None):
r"""Compute :math:`A_0 \mathbf{x}` component of ADMM problem
constraint.
"""
# This calculation involves non-negligible computational cost
# when Xf is None (i.e. the function is not being applied to
# self.X).
if Xf is None:
Xf = sl.rfftn(X, None, self.cri.axisN)
return sl.irfftn(sl.inner(self.Zf, Xf, axis=self.cri.axisM),
self.cri.Nv, self.cri.axisN)
def cbpdn_xstep(k):
"""Do the X step of the cbpdn stage. The only parameter is the slice
index `k` and there are no return values; all inputs and outputs are
from and to global variables.
"""
YU = mp_Z_Y[k] - mp_Z_U[k]
b = mp_DSf[k] + mp_xrho * sl.rfftn(YU, None, mp_cri.axisN)
if mp_cri.Cd == 1:
Xf = sl.solvedbi_sm(mp_Df, mp_xrho, b, axis=mp_cri.axisM)
else:
Xf = sl.solvemdbi_ism(mp_Df, mp_xrho, b, mp_cri.axisM, mp_cri.axisC)
mp_Z_X[k] = sl.irfftn(Xf, mp_cri.Nv, mp_cri.axisN)
def xstep(self):
r"""Minimise Augmented Lagrangian with respect to :math:`\mathbf{x}`.
"""
self.cgit = None
self.YU[:] = self.Y - self.U
b = self.ZSf + self.rho * sl.rfftn(self.YU, None, self.cri.axisN)
self.Xf[:], cgit = sl.solvemdbi_cg(self.Zf, self.rho, b,
self.cri.axisM, self.cri.axisK,
self.opt['CG', 'StopTol'],
self.opt['CG', 'MaxIter'], self.Xf)
self.cgit = cgit
self.X = sl.irfftn(self.Xf, self.cri.Nv, self.cri.axisN)
self.xstep_check(b)
def xstep(self):
r"""Minimise Augmented Lagrangian with respect to :math:`\mathbf{x}`.
"""
self.cgit = None
self.YU[:] = self.Y - self.U
b = self.ZSf + self.rho*sl.rfftn(self.YU, None, self.cri.axisN)
self.Xf[:], cgit = sl.solvemdbi_cg(self.Zf, self.rho, b,
self.cri.axisM, self.cri.axisK,
self.opt['CG', 'StopTol'],
self.opt['CG', 'MaxIter'], self.Xf)
self.cgit = cgit
self.X = sl.irfftn(self.Xf, self.cri.Nv, self.cri.axisN)
self.xstep_check(b)
def xstep(self):
"""Minimise Augmented Lagrangian with respect to :math:`\mathbf{x}`."""
b = self.AHSf + self.rho * np.sum(
np.conj(self.Gf) * sl.rfftn(self.Y - self.U, axes=self.axes),
axis=self.Y.ndim - 1)
self.Xf = b / (self.AHAf + self.rho * self.GHGf)
self.X = sl.irfftn(self.Xf, None, axes=self.axes)
if self.opt['LinSolveCheck']:
ax = (self.AHAf + self.rho * self.GHGf) * self.Xf
self.xrrs = sl.rrs(ax, b)
else:
self.xrrs = None
def xistep(self, i):
r"""Minimise Augmented Lagrangian with respect to :math:`\mathbf{x}`
component :math:`\mathbf{x}_i`.
"""
self.YU[:] = self.Y - self.U[..., i]
b = np.take(self.ZSf, [i], axis=self.cri.axisK) + \
self.rho*sl.rfftn(self.YU, None, self.cri.axisN)
self.Xf[..., i] = sl.solvedbi_sm(np.take(
self.Zf, [i], axis=self.cri.axisK),
self.rho, b, axis=self.cri.axisM)
self.X[..., i] = sl.irfftn(self.Xf[..., i], self.cri.Nv,
self.cri.axisN)
def proximal_step(self, gradf=None):
""" Compute proximal update (gradient descent + constraint)
"""
if gradf is None:
gradf = self.eval_gradf()
Vf = self.Yf - (1. / self.L) * gradf
V = sl.irfftn(Vf, self.cri.Nv, self.cri.axisN)
self.X = self.eval_proxop(V)
self.Xf = sl.rfftn(self.X, None, self.cri.axisN)
def xstep(self):
r"""Minimise Augmented Lagrangian with respect to
:math:`\mathbf{x}`.
"""
self.YU[:] = self.Y - self.U
self.block_sep0(self.YU)[:] += self.S
YUf = sl.rfftn(self.YU, None, self.cri.axisN)
b = sl.inner(np.conj(self.Zf), self.block_sep0(YUf),
axis=self.cri.axisK) + self.block_sep1(YUf)
self.Xf[:] = sl.solvemdbi_ism(self.Zf, 1.0, b, self.cri.axisM,
self.cri.axisK)
self.X = sl.irfftn(self.Xf, self.cri.Nv, self.cri.axisN)
self.xstep_check(b)
def xstep(self):
r"""Minimise Augmented Lagrangian with respect to
:math:`\mathbf{x}`.
"""
self.YU[:] = self.Y - self.U
self.block_sep0(self.YU)[:] += self.S
YUf = sl.rfftn(self.YU, None, self.cri.axisN)
b = sl.inner(np.conj(self.Zf), self.block_sep0(YUf),
axis=self.cri.axisK) + self.block_sep1(YUf)
self.Xf[:] = sl.solvemdbi_ism(self.Zf, 1.0, b, self.cri.axisM,
self.cri.axisK)
self.X = sl.irfftn(self.Xf, self.cri.Nv, self.cri.axisN)
self.xstep_check(b)
def xstep(self):
r"""Minimise Augmented Lagrangian with respect to
:math:`\mathbf{x}`.
"""
b = self.AHSf + self.rho*np.sum(
np.conj(self.Gf)*sl.rfftn(self.Y-self.U, axes=self.axes),
axis=self.Y.ndim-1)
self.Xf = b / (self.AHAf + self.rho*self.GHGf)
self.X = sl.irfftn(self.Xf, self.axsz, axes=self.axes)
if self.opt['LinSolveCheck']:
ax = (self.AHAf + self.rho*self.GHGf)*self.Xf
self.xrrs = sl.rrs(ax, b)
else:
self.xrrs = None
def proximal_step(self, gradf=None):
"""Compute proximal update (gradient descent + constraint).
Variables are mapped back and forth between input and
frequency domains.
"""
if gradf is None:
gradf = self.eval_grad()
self.Vf[:] = self.Yf - (1. / self.L) * gradf
V = sl.irfftn(self.Vf, self.cri.Nv, self.cri.axisN)
self.X[:] = self.eval_proxop(V)
self.Xf = sl.rfftn(self.X, None, self.cri.axisN)
return gradf
def proximal_step(self, gradf=None):
"""Compute proximal update (gradient descent + constraint).
Variables are mapped back and forth between input and
frequency domains.
"""
if gradf is None:
gradf = self.eval_grad()
self.Vf[:] = self.Yf - (1. / self.L) * gradf
V = sl.irfftn(self.Vf, self.cri.Nv, self.cri.axisN)
self.X[:] = self.eval_proxop(V)
self.Xf = sl.rfftn(self.X, None, self.cri.axisN)
return gradf
def xistep(self, i):
r"""Minimise Augmented Lagrangian with respect to :math:`\mathbf{x}`
component :math:`\mathbf{x}_i`.
"""
self.YU[:] = self.Y - self.U[..., i]
b = np.take(self.ZSf, [i], axis=self.cri.axisK) + \
self.rho*sl.rfftn(self.YU, None, self.cri.axisN)
self.Xf[..., i] = sl.solvedbi_sm(np.take(self.Zf, [i],
axis=self.cri.axisK),
self.rho,
b,
axis=self.cri.axisM)
self.X[..., i] = sl.irfftn(self.Xf[..., i], self.cri.Nv,
self.cri.axisN)
def obfn_f(self, Xf=None):
r"""Compute data fidelity term :math:`(1/2) \| W (\sum_m
\mathbf{d}_m * \mathbf{x}_{m} - \mathbf{s}) \|_2^2`.
This is used for backtracking. Since the backtracking is
computed in the DFT, it is important to preserve the
DFT scaling.
"""
if Xf is None:
Xf = self.Xf
Rf = self.eval_Rf(Xf)
R = sl.irfftn(Rf, self.cri.Nv, self.cri.axisN)
WRf = sl.rfftn(self.W * R, self.cri.Nv, self.cri.axisN)
return 0.5 * np.linalg.norm(WRf.flatten(), 2)**2
def obfn_f(self, Xf=None):
r"""Compute data fidelity term :math:`(1/2) \sum_k \| W (\sum_m
\mathbf{d}_m * \mathbf{x}_{k,m} - \mathbf{s}_k) \|_2^2`.
This is used for backtracking. Since the backtracking is
computed in the DFT, it is important to preserve the
DFT scaling.
"""
if Xf is None:
Xf = self.Xf
Rf = self.eval_Rf(Xf)
R = sl.irfftn(Rf, self.cri.Nv, self.cri.axisN)
WRf = sl.rfftn(self.W * R, self.cri.Nv, self.cri.axisN)
return 0.5 * np.linalg.norm(WRf.flatten(), 2)**2
def relax_AX(self):
"""The parent class method that this method overrides only
implements the relaxation step for the variables of the baseline
consensus algorithm. This method calls the overridden method and
then implements the relaxation step for the additional variables
required for the mask decoupling modification to the baseline
algorithm.
"""
super(ConvCnstrMODMaskDcpl_Consensus, self).relax_AX()
self.AX1nr = sl.irfftn(
sl.inner(self.Zf, self.swapaxes(self.Xf), axis=self.cri.axisM),
self.cri.Nv, self.cri.axisN)
if self.rlx == 1.0:
self.AX1 = self.AX1nr
else:
alpha = self.rlx
self.AX1 = alpha * self.AX1nr + (1 - alpha) * (self.Y1 + self.S)
def eval_gradf(self):
"""Compute gradient in Fourier domain."""
# Compute X D - S
self.Ryf = self.eval_Rf(self.Yf)
# Map to spatial domain to multiply by mask
Ry = sl.irfftn(self.Ryf, self.cri.Nv, self.cri.axisN)
# Multiply by mask
WRy = (self.W**2) * Ry
# Map back to frequency domain
WRyf = sl.rfftn(WRy, self.cri.Nv, self.cri.axisN)
gradf = sl.inner(np.conj(self.Zf), WRyf, axis=self.cri.axisK)
# Multiple channel signal, single channel dictionary
if self.cri.C > 1 and self.cri.Cd == 1:
gradf = np.sum(gradf, axis=self.cri.axisC, keepdims=True)
return gradf
def eval_grad(self):
"""Compute gradient in Fourier domain."""
# Compute D X - S
self.Ryf[:] = self.eval_Rf(self.Yf)
# Map to spatial domain to multiply by mask
Ry = sl.irfftn(self.Ryf, self.cri.Nv, self.cri.axisN)
# Multiply by mask
self.WRy[:] = (self.W**2) * Ry
# Map back to frequency domain
WRyf = sl.rfftn(self.WRy, self.cri.Nv, self.cri.axisN)
gradf = np.conj(self.Df) * WRyf
# Multiple channel signal, multiple channel dictionary
if self.cri.Cd > 1:
gradf = np.sum(gradf, axis=self.cri.axisC, keepdims=True)
return gradf
def xstep(self):
r"""Minimise Augmented Lagrangian with respect to
:math:`\mathbf{x}`.
"""
self.cgit = None
self.YU[:] = self.Y - self.U
self.block_sep0(self.YU)[:] += self.S
YUf = sl.rfftn(self.YU, None, self.cri.axisN)
b = sl.inner(np.conj(self.Zf), self.block_sep0(YUf),
axis=self.cri.axisK) + self.block_sep1(YUf)
self.Xf[:], cgit = sl.solvemdbi_cg(self.Zf, 1.0, b,
self.cri.axisM, self.cri.axisK,
self.opt['CG', 'StopTol'],
self.opt['CG', 'MaxIter'], self.Xf)
self.cgit = cgit
self.X = sl.irfftn(self.Xf, self.cri.Nv, self.cri.axisN)
self.xstep_check(b)
def derivD_spdomain(cri, Xr, Sr, Df, Xf, dict_Nv):
B = sl.irfftn(sl.inner(Df, Xf, axis=cri.axisM), s=cri.Nv, axes=cri.axisN) - Sr
B = B[np.newaxis, np.newaxis,]
Xshifted = np.ones(dict_Nv + Xr.shape) * Xr
N1 = 0
N2 = 1
I = 2
J = 3
print("start shifting")
for n1 in range(dict_Nv[0]):
for n2 in range(dict_Nv[1]):
Xshifted[n1][n2] = np.roll(Xshifted[n1][n2], (n1, n2), axis=(I, J))
# print("shifted ", (n1, n2))
ret = np.sum(np.conj(B) * Xshifted, axis=(I, J, 2 + cri.axisK), keepdims=True)
print(ret.shape)
ret = ret[:, :, 0, 0]
print(ret.shape)
return ret
def solve(self):
"""Call the solve method of the inner KConvBPDN object and return the
result.
"""
itst = []
# Main optimisation iterations
for self.j in range(self.j, self.j + self.opt['MaxMainIter']):
for l in range(self.cri.dimN):
# Pre x-step
Wl = self.convolvedict(l) # convolvedict
self.xstep[l].setdictf(Wl) # setdictf
# Solve KCSC
self.xstep[l].solve()
# Post x-step
Kl = np.moveaxis(self.xstep[l].getcoef().squeeze(), [0, 1],
[1, 0])
self.Kf[l] = sl.fftn(Kl, None, [0]) # Update Kruskal
# IterationStats
xitstat = self.xstep.itstat[-1] if self.xstep.itstat else \
self.xstep.IterationStats(
*([0.0,] * len(self.xstep.IterationStats._fields)))
itst += self.isc_lst[l].iterstats(self.j, 0, xitstat,
0) # Accumulate
self.itstat.append(
self.isc(*itst)) # Cast to global itstats and store
# Decomposed ifftn
for l in range(self.cri.dimN):
self.K[l] = sl.irfftn(self.Kf[l], self.cri.Nv[l],
[0]) # ifft transform
self.j += 1
def xstep(self):
r"""Minimise Augmented Lagrangian with respect to
:math:`\mathbf{x}`."""
self.YU[:] = self.Y - self.U
Zf = sl.rfftn(self.YU, None, self.cri.axisN)
ZfQ = sl.dot(self.Q.T, Zf, axis=self.cri.axisC)
b = self.DSfBQ + self.rho * ZfQ
Xh = sl.solvedbi_sm(self.gDf, self.rho, b, self.c, axis=self.cri.axisM)
self.Xf[:] = sl.dot(self.Q, Xh, axis=self.cri.axisC)
self.X = sl.irfftn(self.Xf, self.cri.Nv, self.cri.axisN)
if self.opt['LinSolveCheck']:
DDXf = np.conj(self.Df) * sl.inner(
self.Df, self.Xf, axis=self.cri.axisM)
DDXfBB = sl.dot(self.B.T.dot(self.B), DDXf, axis=self.cri.axisC)
ax = DDXfBB + self.rho * self.Xf
b = sl.dot(self.B.T, self.DSf, axis=self.cri.axisC) + \
self.rho * Zf
self.xrrs = sl.rrs(ax, b)
else:
self.xrrs = None
def evaluate(self):
"""Evaluate functional value of previous iteration."""
if self.opt['AccurateDFid']:
DX = self.reconstruct()
W = self.dstep.W
S = self.dstep.S
else:
W = mp_W
S = mp_S
Xf = mp_Zf
Df = mp_Df
DX = sl.irfftn(
sl.inner(Df[np.newaxis, ...],
Xf,
axis=self.xstep.cri.axisM + 1), self.xstep.cri.Nv,
np.array(self.xstep.cri.axisN) + 1)
dfd = (np.linalg.norm(W * (DX - S))**2) / 2.0
rl1 = np.sum(np.abs(self.getcoef()))
obj = dfd + self.xstep.lmbda * rl1
return (obj, dfd, rl1)
def xstep(self):
r"""Minimise Augmented Lagrangian with respect to block vector
:math:`\mathbf{x} = \left( \begin{array}{ccc} \mathbf{x}_0^T &
\mathbf{x}_1^T & \ldots \end{array} \right)^T\;`.
"""
# This test reflects empirical evidence that two slightly
# different implementations are faster for single or
# multi-channel data. This kludge is intended to be temporary.
if self.cri.Cd > 1:
for i in range(self.Nb):
self.xistep(i)
else:
self.YU[:] = self.Y[..., np.newaxis] - self.U
b = np.swapaxes(self.ZSf[..., np.newaxis], self.cri.axisK, -1) \
+ self.rho*sl.rfftn(self.YU, None, self.cri.axisN)
for i in range(self.Nb):
self.Xf[..., i] = sl.solvedbi_sm(self.Zf[..., [i], :],
self.rho,
b[..., i],
axis=self.cri.axisM)
self.X = sl.irfftn(self.Xf, self.cri.Nv, self.cri.axisN)
if self.opt['LinSolveCheck']:
ZSfs = np.sum(self.ZSf, axis=self.cri.axisK, keepdims=True)
YU = np.sum(self.Y[..., np.newaxis] - self.U, axis=-1)
b = ZSfs + self.rho * sl.rfftn(YU, None, self.cri.axisN)
Xf = self.swapaxes(self.Xf)
Zop = lambda x: sl.inner(self.Zf, x, axis=self.cri.axisM)
ZHop = lambda x: np.conj(self.Zf) * x
ax = np.sum(ZHop(Zop(Xf)) + self.rho * Xf,
axis=self.cri.axisK,
keepdims=True)
self.xrrs = sl.rrs(ax, b)
else:
self.xrrs = None
def dstep(self):
"""Compute dictionary update for training data of preceding
:meth:`xstep`.
"""
# Compute X D - S
Ryf = sl.inner(self.Zf, self.Df, axis=self.cri.axisM) - self.Sf
# Compute gradient
gradf = sl.inner(np.conj(self.Zf), Ryf, axis=self.cri.axisK)
# If multiple channel signal, single channel dictionary
if self.cri.C > 1 and self.cri.Cd == 1:
gradf = np.sum(gradf, axis=self.cri.axisC, keepdims=True)
# Update gradient step
self.eta = self.eta_a / (self.j + self.eta_b)
# Compute gradient descent
self.Gf[:] = self.Df - self.eta * gradf
self.G = sl.irfftn(self.Gf, self.cri.Nv, self.cri.axisN)
# Eval proximal operator
self.Dprv[:] = self.D
self.D[:] = self.Pcn(self.G)
def xstep(self):
r"""Minimise Augmented Lagrangian with respect to
:math:`\mathbf{x}`."""
self.YU[:] = self.Y - self.U
YUf = sl.rfftn(self.YU, None, self.cri.axisN)
YUf0 = self.block_sep0(YUf)
YUf1 = self.block_sep1(YUf)
b = self.rho * np.sum(np.conj(self.GDf) * YUf1, axis=-1)
if self.cri.Cd > 1:
b = np.sum(b, axis=self.cri.axisC, keepdims=True)
b += self.DSf + self.rho * YUf0
# Concatenate multiple GDf components on axisC. For
# single-channel signals, and multi-channel signals with a
# single-channel dictionary, we end up with sl.solvemdbi_ism
# solving a linear system of rank dimN+1 (corresponding to the
# dictionary and a gradient operator per spatial dimension) plus
# an identity. For multi-channel signals with a multi-channel
# dictionary, we end up with sl.solvemdbi_ism solving a linear
# system of rank C.d (dimN+1) (corresponding to the dictionary
# and a gradient operator per spatial dimension for each
# channel) plus an identity.
# The structure of the linear system to be solved depends on the
# number of channels in the signal and dictionary. Both branches are
# the same in the single-channel signal case (the choice of handling
# it via the 'else' branch is somewhat arbitrary).
if self.cri.C > 1 and self.cri.Cd == 1:
# Concatenate multiple GDf components on the final axis
# of GDf (that indexes the number of gradient operators). For
# multi-channel signals with a single-channel dictionary,
# sl.solvemdbi_ism has to solve a linear system of rank dimN+1
# (corresponding to the dictionary and a gradient operator per
# spatial dimension)
DfGDf = np.concatenate([
self.Df[..., np.newaxis],
] + [
np.sqrt(self.rho) * self.GDf[..., k, np.newaxis]
for k in range(self.GDf.shape[-1])
],
axis=-1)
self.Xf[:] = sl.solvemdbi_ism(DfGDf, self.rho, b[..., np.newaxis],
self.cri.axisM, -1)[..., 0]
else:
# Concatenate multiple GDf components on axisC. For multi-channel
# signals with a multi-channel dictionary, sl.solvemdbi_ism has
# to solve a linear system of rank C.d (dimN+1) (corresponding to
# the dictionary and a gradient operator per spatial dimension
# for each channel) plus an identity.
DfGDf = np.concatenate([
self.Df,
] + [
np.sqrt(self.rho) * self.GDf[..., k]
for k in range(self.GDf.shape[-1])
],
axis=self.cri.axisC)
self.Xf[:] = sl.solvemdbi_ism(DfGDf, self.rho, b, self.cri.axisM,
self.cri.axisC)
self.X = sl.irfftn(self.Xf, self.cri.Nv, self.cri.axisN)
if self.opt['LinSolveCheck']:
if self.cri.C > 1 and self.cri.Cd == 1:
Dop = lambda x: sl.inner(
DfGDf, x[..., np.newaxis], axis=self.cri.axisM)
DHop = lambda x: sl.inner(np.conj(DfGDf), x, axis=-1)
ax = DHop(Dop(self.Xf))[..., 0] + self.rho * self.Xf
else:
Dop = lambda x: sl.inner(DfGDf, x, axis=self.cri.axisM)
DHop = lambda x: sl.inner(
np.conj(DfGDf), x, axis=self.cri.axisC)
ax = DHop(Dop(self.Xf)) + self.rho * self.Xf
self.xrrs = sl.rrs(ax, b)
else:
self.xrrs = None
def mysolve(cri,
Dr0,
Xr,
Sr,
final_sigma,
maxitr=40,
param_mu=1,
debug_dir=None):
Dr = Dr0.copy()
Xr = Xr.copy()
Sr = Sr.copy()
#離散フーリエ変換
Df = sl.rfftn(Dr, s=cri.Nv, axes=cri.axisN)
Sf = sl.rfftn(Sr, s=cri.Nv, axes=cri.axisN)
Xf = sl.rfftn(Xr, s=cri.Nv, axes=cri.axisN)
alpha = 1e0
# sigma set
first_sigma = Xr.max() * 4
# σ←cσ(c < 1)のｃの決定
c = (final_sigma / first_sigma)**(1 / (maxitr - 1))
print("c = %.8f" % c)
sigma_list = []
sigma_list.append(first_sigma)
for i in range(maxitr - 1):
sigma_list.append(sigma_list[i] * c)
print(sigma_list[-1])
# 辞書のクロップする領域を添え字で指定
crop_op = []
for l in Dr.shape:
crop_op.append(slice(0, l))
crop_op = tuple(crop_op)
print(crop_op)
updcnt = 0
for sigma in sigma_list:
print("sigma = %.8f" % sigma)
# print("l0norm: %f" % l0norm(Xr, sigma_list[-1]))
# print('error1: ', l2norm(Sr - reconstruct(cri, Dr, Xr)))
delta = Xr * np.exp(-(Xr * Xr) / (2 * sigma * sigma))
# print("l2(Xr): %.6f, l2(delta): %.6f" % (l2norm(Xr), l2norm(delta)))
Xr = Xr - param_mu * delta # + np.random.randn(*Xr.shape)*sigma*1e-1
Xf = sl.rfftn(Xr, cri.Nv, cri.axisN)
# saveXhist(Xr, "./hist/%db.png" % reccnt)
# print('error2: ', l2norm(Sr - reconstruct(cri, Dr, Xr)))
# if debug_dir is not None:
# save_reconstructed(cri, Dr, Xr, Sr, debug_dir + '/%drecA.png' % updcnt)
# DXf = sl.inner(Df, Xf, axis=cri.axisM)
# gamma = (np.sum(np.conj(DXf) * Sf, axis=cri.axisN, keepdims=True) + np.sum(DXf * np.conj(Sf), axis=cri.axisN, keepdims=True)) / 2 / np.sum(np.conj(DXf) * DXf, axis=cri.axisN, keepdims=True)
# print(gamma)
# print(gamma.shape, ' * ', Xr.shape)
# gamma = np.real(gamma)
# Xr = Xr * gamma
# Xf = to_frequency(cri, Xr)
# if debug_dir is not None:
# save_reconstructed(cri, Dr, Xr, Sr, debug_dir + '/%drecB.png' % updcnt)
# print('error3: ', l2norm(Sr - reconstruct(cri, Dr, Xr)))
# print("max: ", np.max(Xr))
# 辞書の勾配降下
B = sl.inner(Xf, Df, axis=cri.axisM) - Sf
derivDf = sl.inner(np.conj(Xf), B,
axis=cri.axisK) # 目的関数(信号との二乗近似誤差)の勾配
# derivDr = sl.irfftn(derivDf, s=cri.Nv, axes=cri.axisN)[crop_op]
def func(alpha):
Df_ = Df - alpha * derivDf
Dr_ = sl.irfftn(Df_, s=cri.Nv, axes=cri.axisN)[crop_op]
Df_ = sl.rfftn(Dr_, s=cri.Nv, axes=cri.axisN)
Sf_ = sl.inner(Df_, Xf, axis=cri.axisM)
return l2norm(Sr - sl.irfftn(Sf_, s=cri.Nv, axes=cri.axisN))
choice = np.array([func(alpha / 2),
func(alpha),
func(alpha * 2)]).argmin()
alpha *= [0.5, 1, 2][choice]
print("alpha: ", alpha)
Df = Df - alpha * derivDf
# 辞書の射影
Dr = sl.irfftn(Df, s=cri.Nv, axes=cri.axisN)
Pcn = cnvrep.getPcn(Dr.shape, cri.Nv, cri.dimN, cri.dimCd,
zm=False) # 射影関数のインスタンス化
Dr = Pcn(Dr)
Dr = Dr[crop_op]
print(l2norm(Dr.T[0]))
Df = sl.rfftn(Dr, s=cri.Nv, axes=cri.axisN)
b = sl.inner(Df, Xf, axis=cri.axisM) - Sf
c = sl.inner(Df, np.conj(Df), axis=cri.axisM)
Xf = Xf - np.conj(Df) / c * b
Xr = sl.irfftn(Xf, s=cri.Nv, axes=cri.axisN)
# save_reconstructed(cri, Dr, Xr, Sr, debug_dir + "rec/%dd.png" % updcnt)
# saveXhist(Xr, debug_dir + "hist/%db.png" % updcnt)
updcnt += 1
# print("l0 norm of final X: %d" % smoothedl0norm(Xr, 0.00001))
plot.close()
mplot.close()
return Dr
def func(alpha):
Df_ = Df - alpha * derivDf
Dr_ = sl.irfftn(Df_, s=cri.Nv, axes=cri.axisN)[crop_op]
Df_ = sl.rfftn(Dr_, s=cri.Nv, axes=cri.axisN)
Sf_ = sl.inner(Df_, Xf, axis=cri.axisM)
return l2norm(Sr - sl.irfftn(Sf_, s=cri.Nv, axes=cri.axisN))
def convert_to_X(Xf, cri):
X = sl.irfftn(Xf, cri.Nv, cri.axisN).squeeze()
return X
def convert_to_S(Sf, cri):
S = sl.irfftn(Sf, cri.Nv, cri.axisN).squeeze()
return S
def convert_to_D(Df, dsz, cri):
D = sl.irfftn(Df, cri.Nv, cri.axisN)
D = cr.bcrop(D, dsz, cri.dimN).squeeze()
return D
# gray=True, idxexp=np.s_[:, 160:672])
img = mpimg.imread('barbara1.png')
np.random.seed(12345)
imgn = img + np.random.normal(0.0, 0.1, img.shape)
print("Noisy image PSNR: %5.2f dB" % sm.psnr(img, imgn))
npd = 16
fltlmbd = 5.0
imgnl, imgnh = util.tikhonov_filter(imgn, fltlmbd, npd)
D = util.convdicts()['G:8x8x32']
D = D[:, :, 0:14]
# D = np.random.randn(8, 8, 14)
imgnpl, imgnph = util.tikhonov_filter(pad(imgn), fltlmbd, npd)
W = spl.irfftn(
np.conj(spl.rfftn(D, imgnph.shape,
(0, 1))) * spl.rfftn(imgnph[..., np.newaxis], None,
(0, 1)), imgnph.shape, (0, 1))
W = W**2
W = 1.0 / (np.maximum(np.abs(W), 1e-8))
lmbda = 1.5e-2
mu = 0.005
opt1 = cbpdn.ConvBPDN.Options({
'Verbose': True,
'MaxMainIter': 250,
'HighMemSolve': True,
'RelStopTol': 3e-3,
'AuxVarObj': True,
'L1Weight': W,
'AutoRho': {
'Enabled': False
def reconstruct(cri, Dr, Xr):
Xf = sl.rfftn(Xr, s=cri.Nv, axes=cri.axisN)
Df = sl.rfftn(Dr, s=cri.Nv, axes=cri.axisN)
return sl.irfftn(sl.inner(Df, Xf, axis=cri.axisM),
s=cri.Nv,
axes=cri.axisN)
def convDictLearn(cri,
Dr0,
dsz,
Sr,
final_sigma,
maxitr,
non_nega,
param_mu=1,
debug_dir=None):
Dr = Dr0.copy()
Sr = Sr.copy()
# 係数をl2ノルム最小解で初期化
Xr = l2norm_minimize(cri, Dr, Sr)
# 2次元離散フーリエ変換
Df = sl.rfftn(Dr, s=cri.Nv, axes=cri.axisN)
Sf = sl.rfftn(Sr, s=cri.Nv, axes=cri.axisN)
Xf = sl.rfftn(Xr, s=cri.Nv, axes=cri.axisN)
alpha = 1e0
# sigma set
first_sigma = Xr.max() * 4
# σを更新する定数c(c<1)の決定
c = (final_sigma / first_sigma)**(1 / (maxitr - 1))
print("c = %.8f" % c)
sigma_list = []
sigma_list.append(first_sigma)
for i in range(maxitr - 1):
sigma_list.append(sigma_list[i] * c)
# 辞書のクロップする領域を添え字で指定
crop_op = []
for l in Dr.shape:
crop_op.append(slice(0, l))
crop_op = tuple(crop_op)
# 射影関数のインスタンス化
Pcn = cnvrep.getPcn(dsz, cri.Nv, cri.dimN, cri.dimCd, zm=False)
updcnt = 0
for sigma in sigma_list:
print("sigma = %.8f" % sigma, end=" ")
# 係数の勾配降下
delta = Xr * np.exp(-(Xr * Xr) / (2 * sigma * sigma))
Xr = Xr - param_mu * delta
Xf = sl.rfftn(Xr, cri.Nv, cri.axisN)
print("l0norm = %i" % np.where(
abs(Xr.transpose(3, 4, 2, 0, 1).squeeze()[0]) < final_sigma, 0,
1).sum(),
end=" ")
# 辞書の勾配降下
B = sl.inner(Xf, Df, axis=cri.axisM) - Sf
derivDf = sl.inner(np.conj(Xf), B, axis=cri.axisK)
def func(alpha):
Df_ = Df - alpha * derivDf
Dr_ = sl.irfftn(Df_, s=cri.Nv, axes=cri.axisN)[crop_op]
Df_ = sl.rfftn(Dr_, s=cri.Nv, axes=cri.axisN)
Sf_ = sl.inner(Df_, Xf, axis=cri.axisM)
return myutil.l2norm(Sr - sl.irfftn(Sf_, s=cri.Nv, axes=cri.axisN))
error_list = np.array([func(alpha / 2), func(alpha), func(alpha * 2)])
choice = error_list.argmin()
alpha *= [0.5, 1, 2][choice]
print("alpha = %.8f" % alpha, end=" ")
print("error = %5.5f" % error_list[choice], end=" ")
Df = Df - alpha * derivDf
# 辞書の射影
Dr = Pcn(sl.irfftn(Df, s=cri.Nv,
axes=cri.axisN))[crop_op] # 正規化とゼロパディングを同時に行う
# print(myutil.l2norm(Dr.T[0]))
# 係数の射影
Df = sl.rfftn(Dr, s=cri.Nv, axes=cri.axisN)
b = sl.inner(Df, Xf, axis=cri.axisM) - Sf
c = sl.inner(Df, np.conj(Df), axis=cri.axisM)
Xf = Xf - np.conj(Df) / c * b
Xr = sl.irfftn(Xf, s=cri.Nv, axes=cri.axisN)
if (non_nega): Xr = np.where(Xr < 0, 0, Xr) # 非負制約
# Xr = np.where(Xr < 1e-6, 0, Xr)
print("l0norm_projected = %i" % np.where(
abs(Xr.transpose(3, 4, 2, 0, 1).squeeze()[0]) < final_sigma, 0,
1).sum())
updcnt += 1
return Dr, Xr
def to_spatial(cri, Af):
return sl.irfftn(Af, s=cri.Nv, axes=cri.axisN)