def test_09(self):
rho = 1e-1
N = 32
M = 16
K = 8
D = complex_randn(N, N, 1, 1, M)
X = complex_randn(N, N, 1, K, M)
S = np.sum(D*X, axis=4, keepdims=True)
Z = (D.conj()*np.sum(D*X, axis=4, keepdims=True) + \
rho*X - D.conj()*S) / rho
Xslv = linalg.solvedbi_sm(D, rho, D.conj()*S + rho*Z)
assert linalg.rrs(D.conj()*np.sum(D*Xslv, axis=4, keepdims=True) +
rho*Xslv, D.conj()*S + rho*Z) < 1e-11
def test_15(self):
rho = 1e-1
N = 32
M = 16
K = 8
D = complex_randn(N, N, 1, 1, M)
X = complex_randn(N, N, 1, K, M)
S = np.sum(D*X, axis=4, keepdims=True)
Xop = lambda x: np.sum(X * x, axis=4, keepdims=True)
XHop = lambda x: np.sum(np.conj(X) * x, axis=3, keepdims=True)
Z = (XHop(Xop(D)) + rho*D - XHop(S)) / rho
Dslv, cgit = linalg.solvemdbi_cg(X, rho, XHop(S)+rho*Z, 4, 3, tol=1e-6)
assert linalg.rrs(XHop(Xop(Dslv)) + rho*Dslv, XHop(S) + rho*Z) <= 1e-6
def test_10(self):
N = 32
M = 16
K = 8
D = complex_randn(N, N, 1, 1, M)
X = complex_randn(N, N, 1, K, M)
S = np.sum(D*X, axis=4, keepdims=True)
d = 1e-1 * (np.random.randn(N, N, 1, 1, M).astype('complex') +
np.random.randn(N, N, 1, 1, M).astype('complex') * 1.0j)
Z = (D.conj()*np.sum(D*X, axis=4, keepdims=True) +
d*X - D.conj()*S) / d
Xslv = linalg.solvedbd_sm(D, d, D.conj()*S + d*Z)
assert linalg.rrs(D.conj()*np.sum(D*Xslv, axis=4, keepdims=True) +
d*Xslv, D.conj()*S + d*Z) < 1e-11
def test_14(self):
rho = 1e-1
N = 64
M = 32
C = 3
K = 8
D = complex_randn(N, N, C, 1, M)
X = complex_randn(N, N, 1, K, M)
S = np.sum(D*X, axis=4, keepdims=True)
Xop = lambda x: np.sum(X * x, axis=4, keepdims=True)
XHop = lambda x: np.sum(np.conj(X) * x, axis=3, keepdims=True)
Z = (XHop(Xop(D)) + rho*D - XHop(S)) / rho
Dslv = linalg.solvemdbi_rsm(X, rho, XHop(S) + rho*Z, 3)
assert linalg.rrs(XHop(Xop(Dslv)) + rho*Dslv, XHop(S) + rho*Z) < 1e-11
t = np.linspace(-2, 2, N, endpoint=False)
#the first element is the length of signal, the fourth one is how
s_noise = np.zeros((N,K), dtype=complex)
s_clean = np.zeros((N,K), dtype=complex)
Dtemp = np.zeros((N,K), dtype=complex)
# delta matrix to shift the Gaussian pulses
delta = np.zeros((N, K), dtype=complex)
# construct sine/cosine signal
for i in range(K):
sigma = 0.05 # noise level
noise = sigma * si.complex_randn(N)
delta[np.random.randint(0, 32), i] = 1
real,imag,e = signal.gausspulse(t, 5, retquad=True, retenv=True)
temp = real + complex(0, 1) * imag
temp = np.convolve(temp,delta[:,i],'same')
s_clean[:,i] = temp
s_noise[:,i] = s_clean[:,i] + noise
Dtemp[:,i] = s_clean[:,i] + 10*noise
D0 =Dtemp[:,0:M]
# Function computing reconstruction error at lmbda
Maxiter = 400
opt_par = comcbpdn.ComplexConvBPDN.Options({'FastSolve': True, 'Verbose': True, 'StatusHeader': False,
'MaxMainIter': Maxiter,'RelStopTol': 5e-5, 'AuxVarObj': True,
# onion_complex = pickle.load(f)
# f.close()
np.seterr(divide='ignore')
N = 128 # signal length
A = 5 # amplitude
M = 10 # filter number
fd = np.linspace(5, 20, M)
fs = 128 # sampling frequency must be higher enough to eliminate cutoff
dimN = 1 # spatial dimension 1
# create a complex gaussian noise using sporco.signal.complex_randn,
# same length as input signal
noise = si.complex_randn(N)
D0 = np.zeros((N, M), dtype=complex)
# construct a complex sine dictionary
for i in range(len(fd)):
D0[:, i] = sinesignal.generate_sine_wave_no(N, A, fd[i], fs, 0) + \
complex(0, 1) * sinesignal.generate_sine_wave_no(N, A, fd[i], fs, 0)
lmbda_0 = M * abs(noise.max())
# construct a 5Hz sine signal
s_clean = np.zeros(N)
s_clean = sinesignal.generate_sine_wave_no(N, A, 5, fs, 0) \
+ complex(0, 1) * sinesignal.generate_sine_wave_no(N, A, 5, fs, 0)
