Parallel Single-channel CSC¶
This example compares the use of parcbpdn.ParConvBPDN with
admm.cbpdn.ConvBPDN solving a convolutional sparse coding
problem with a greyscale signal
\[\mathrm{argmin}_\mathbf{x} \; \frac{1}{2} \left\| \sum_m \mathbf{d}_m * \mathbf{x}_{m} - \mathbf{s} \right\|_2^2 + \lambda \sum_m \| \mathbf{x}_{m} \|_1 \;,\]
where \(\mathbf{d}_{m}\) is the \(m^{\text{th}}\) dictionary filter, \(\mathbf{x}_{m}\) is the coefficient map corresponding to the \(m^{\text{th}}\) dictionary filter, and \(\mathbf{s}\) is the input image.
from __future__ import print_function
from builtins import input
import pyfftw # See https://github.com/pyFFTW/pyFFTW/issues/40
import numpy as np
from sporco import util
from sporco import signal
from sporco import plot
plot.config_notebook_plotting()
import sporco.metric as sm
from sporco.admm import cbpdn
from sporco.admm import parcbpdn
Load example image.
img = util.ExampleImages().image('kodim23.png', zoom=1.0, scaled=True,
gray=True, idxexp=np.s_[160:416, 60:316])
Highpass filter example image.
npd = 16
fltlmbd = 10
sl, sh = signal.tikhonov_filter(img, fltlmbd, npd)
Load dictionary and display it.
D = util.convdicts()['G:12x12x216']
plot.imview(util.tiledict(D), fgsz=(7, 7))
lmbda = 5e-2
The RelStopTol option was chosen for the two different methods to stop with similar functional values
Initialise and run standard serial CSC solver using ADMM with an equality constraint [49].
opt = cbpdn.ConvBPDN.Options({'Verbose': True, 'MaxMainIter': 200,
'RelStopTol': 5e-3, 'AuxVarObj': False,
'AutoRho': {'Enabled': False}})
b = cbpdn.ConvBPDN(D, sh, lmbda, opt=opt, dimK=0)
X = b.solve()
Itn Fnc DFid Regℓ1 r s
------------------------------------------------------
0 8.14e+01 3.14e-02 1.63e+03 1.00e+00 0.00e+00
1 7.77e+01 2.13e-01 1.55e+03 9.98e-01 9.36e-03
2 7.51e+01 4.97e-01 1.49e+03 9.79e-01 4.81e-02
3 7.15e+01 8.36e-01 1.41e+03 8.35e-01 1.01e-01
4 6.66e+01 1.14e+00 1.31e+03 4.32e-01 1.29e-01
5 6.62e+01 1.36e+00 1.30e+03 2.80e-01 1.23e-01
6 6.80e+01 1.49e+00 1.33e+03 2.23e-01 1.04e-01
7 6.73e+01 1.58e+00 1.31e+03 1.94e-01 9.24e-02
8 6.43e+01 1.67e+00 1.25e+03 1.75e-01 8.29e-02
9 6.05e+01 1.78e+00 1.17e+03 1.61e-01 7.29e-02
10 5.67e+01 1.91e+00 1.10e+03 1.50e-01 6.27e-02
11 5.35e+01 2.05e+00 1.03e+03 1.39e-01 5.47e-02
12 5.12e+01 2.20e+00 9.80e+02 1.27e-01 4.94e-02
13 4.96e+01 2.34e+00 9.45e+02 1.14e-01 4.60e-02
14 4.88e+01 2.46e+00 9.26e+02 1.02e-01 4.27e-02
15 4.84e+01 2.56e+00 9.17e+02 9.25e-02 3.95e-02
16 4.80e+01 2.64e+00 9.07e+02 8.45e-02 3.72e-02
17 4.71e+01 2.70e+00 8.89e+02 7.80e-02 3.56e-02
18 4.62e+01 2.76e+00 8.69e+02 7.25e-02 3.43e-02
19 4.54e+01 2.81e+00 8.51e+02 6.81e-02 3.21e-02
20 4.44e+01 2.86e+00 8.31e+02 6.44e-02 2.97e-02
21 4.34e+01 2.91e+00 8.11e+02 6.10e-02 2.75e-02
22 4.23e+01 2.96e+00 7.87e+02 5.76e-02 2.57e-02
23 4.12e+01 3.02e+00 7.64e+02 5.45e-02 2.43e-02
24 4.01e+01 3.07e+00 7.41e+02 5.14e-02 2.33e-02
25 3.93e+01 3.13e+00 7.23e+02 4.84e-02 2.24e-02
26 3.87e+01 3.19e+00 7.11e+02 4.58e-02 2.14e-02
27 3.84e+01 3.24e+00 7.03e+02 4.35e-02 2.02e-02
28 3.81e+01 3.28e+00 6.96e+02 4.14e-02 1.92e-02
29 3.78e+01 3.32e+00 6.89e+02 3.95e-02 1.83e-02
30 3.74e+01 3.36e+00 6.80e+02 3.77e-02 1.78e-02
31 3.69e+01 3.39e+00 6.70e+02 3.60e-02 1.74e-02
32 3.65e+01 3.42e+00 6.61e+02 3.45e-02 1.69e-02
33 3.62e+01 3.44e+00 6.54e+02 3.31e-02 1.62e-02
34 3.59e+01 3.47e+00 6.48e+02 3.19e-02 1.55e-02
35 3.56e+01 3.49e+00 6.42e+02 3.08e-02 1.48e-02
36 3.52e+01 3.51e+00 6.34e+02 2.97e-02 1.42e-02
37 3.48e+01 3.53e+00 6.26e+02 2.87e-02 1.37e-02
38 3.43e+01 3.55e+00 6.15e+02 2.76e-02 1.33e-02
39 3.38e+01 3.57e+00 6.05e+02 2.65e-02 1.30e-02
40 3.34e+01 3.60e+00 5.96e+02 2.55e-02 1.27e-02
41 3.30e+01 3.62e+00 5.88e+02 2.46e-02 1.22e-02
42 3.28e+01 3.64e+00 5.82e+02 2.37e-02 1.17e-02
43 3.25e+01 3.66e+00 5.77e+02 2.29e-02 1.13e-02
44 3.23e+01 3.69e+00 5.71e+02 2.22e-02 1.08e-02
45 3.20e+01 3.71e+00 5.66e+02 2.15e-02 1.05e-02
46 3.17e+01 3.73e+00 5.60e+02 2.09e-02 1.03e-02
47 3.14e+01 3.75e+00 5.54e+02 2.02e-02 1.01e-02
48 3.12e+01 3.77e+00 5.49e+02 1.96e-02 9.83e-03
49 3.10e+01 3.78e+00 5.44e+02 1.91e-02 9.54e-03
50 3.08e+01 3.80e+00 5.41e+02 1.86e-02 9.21e-03
51 3.07e+01 3.81e+00 5.38e+02 1.81e-02 8.90e-03
52 3.05e+01 3.83e+00 5.34e+02 1.76e-02 8.66e-03
53 3.03e+01 3.84e+00 5.29e+02 1.71e-02 8.48e-03
54 3.01e+01 3.85e+00 5.25e+02 1.66e-02 8.33e-03
55 2.99e+01 3.87e+00 5.21e+02 1.62e-02 8.19e-03
56 2.98e+01 3.88e+00 5.17e+02 1.57e-02 8.01e-03
57 2.96e+01 3.89e+00 5.14e+02 1.53e-02 7.79e-03
58 2.95e+01 3.90e+00 5.12e+02 1.49e-02 7.55e-03
59 2.94e+01 3.92e+00 5.09e+02 1.46e-02 7.33e-03
60 2.93e+01 3.93e+00 5.07e+02 1.42e-02 7.13e-03
61 2.91e+01 3.94e+00 5.04e+02 1.39e-02 6.97e-03
62 2.90e+01 3.95e+00 5.01e+02 1.35e-02 6.82e-03
63 2.89e+01 3.96e+00 4.98e+02 1.32e-02 6.69e-03
64 2.87e+01 3.97e+00 4.95e+02 1.29e-02 6.52e-03
65 2.86e+01 3.98e+00 4.92e+02 1.26e-02 6.35e-03
66 2.85e+01 3.99e+00 4.90e+02 1.24e-02 6.20e-03
67 2.83e+01 4.00e+00 4.87e+02 1.21e-02 6.08e-03
68 2.82e+01 4.00e+00 4.84e+02 1.18e-02 5.99e-03
69 2.81e+01 4.01e+00 4.81e+02 1.16e-02 5.91e-03
70 2.79e+01 4.02e+00 4.78e+02 1.13e-02 5.83e-03
71 2.78e+01 4.03e+00 4.76e+02 1.11e-02 5.72e-03
72 2.77e+01 4.04e+00 4.74e+02 1.08e-02 5.59e-03
73 2.77e+01 4.04e+00 4.72e+02 1.06e-02 5.45e-03
74 2.76e+01 4.05e+00 4.71e+02 1.04e-02 5.31e-03
75 2.75e+01 4.06e+00 4.70e+02 1.02e-02 5.16e-03
76 2.74e+01 4.06e+00 4.68e+02 9.98e-03 5.05e-03
77 2.73e+01 4.07e+00 4.66e+02 9.78e-03 4.96e-03
78 2.72e+01 4.08e+00 4.63e+02 9.58e-03 4.88e-03
79 2.71e+01 4.08e+00 4.61e+02 9.38e-03 4.81e-03
80 2.70e+01 4.09e+00 4.59e+02 9.19e-03 4.73e-03
81 2.69e+01 4.09e+00 4.57e+02 9.01e-03 4.64e-03
82 2.69e+01 4.10e+00 4.55e+02 8.83e-03 4.53e-03
83 2.68e+01 4.10e+00 4.53e+02 8.66e-03 4.43e-03
84 2.67e+01 4.11e+00 4.52e+02 8.49e-03 4.36e-03
85 2.66e+01 4.11e+00 4.50e+02 8.32e-03 4.27e-03
86 2.65e+01 4.12e+00 4.48e+02 8.16e-03 4.21e-03
87 2.64e+01 4.12e+00 4.46e+02 8.00e-03 4.13e-03
88 2.64e+01 4.13e+00 4.45e+02 7.85e-03 4.06e-03
89 2.63e+01 4.13e+00 4.43e+02 7.71e-03 3.99e-03
90 2.62e+01 4.14e+00 4.42e+02 7.57e-03 3.91e-03
91 2.61e+01 4.14e+00 4.40e+02 7.43e-03 3.83e-03
92 2.61e+01 4.15e+00 4.38e+02 7.30e-03 3.77e-03
93 2.60e+01 4.15e+00 4.37e+02 7.16e-03 3.72e-03
94 2.59e+01 4.16e+00 4.35e+02 7.03e-03 3.67e-03
95 2.59e+01 4.16e+00 4.34e+02 6.91e-03 3.61e-03
96 2.58e+01 4.17e+00 4.33e+02 6.78e-03 3.56e-03
97 2.57e+01 4.17e+00 4.31e+02 6.67e-03 3.50e-03
98 2.57e+01 4.18e+00 4.30e+02 6.55e-03 3.42e-03
99 2.57e+01 4.18e+00 4.30e+02 6.44e-03 3.35e-03
100 2.56e+01 4.18e+00 4.29e+02 6.33e-03 3.28e-03
101 2.56e+01 4.19e+00 4.28e+02 6.22e-03 3.23e-03
102 2.55e+01 4.19e+00 4.27e+02 6.11e-03 3.18e-03
103 2.55e+01 4.19e+00 4.26e+02 6.01e-03 3.14e-03
104 2.54e+01 4.20e+00 4.25e+02 5.91e-03 3.09e-03
105 2.54e+01 4.20e+00 4.23e+02 5.81e-03 3.05e-03
106 2.53e+01 4.20e+00 4.22e+02 5.71e-03 3.00e-03
107 2.53e+01 4.21e+00 4.21e+02 5.62e-03 2.95e-03
108 2.52e+01 4.21e+00 4.20e+02 5.52e-03 2.90e-03
109 2.52e+01 4.21e+00 4.19e+02 5.43e-03 2.85e-03
110 2.51e+01 4.21e+00 4.18e+02 5.34e-03 2.81e-03
111 2.51e+01 4.22e+00 4.17e+02 5.25e-03 2.76e-03
112 2.50e+01 4.22e+00 4.16e+02 5.17e-03 2.71e-03
113 2.50e+01 4.22e+00 4.15e+02 5.09e-03 2.65e-03
114 2.50e+01 4.23e+00 4.15e+02 5.01e-03 2.59e-03
115 2.49e+01 4.23e+00 4.14e+02 4.94e-03 2.54e-03
------------------------------------------------------
Initialise and run parallel CSC solver using ADMM dictionary partition method [42].
opt_par = parcbpdn.ParConvBPDN.Options({'Verbose': True, 'MaxMainIter': 200,
'RelStopTol': 1e-2, 'AuxVarObj': False, 'AutoRho':
{'Enabled': False}, 'alpha': 2.5})
b_par = parcbpdn.ParConvBPDN(D, sh, lmbda, opt=opt_par, dimK=0)
X_par = b_par.solve()
Itn Fnc DFid Regl1 r s
------------------------------------------------------
0 4.77e+02 2.95e+02 3.65e+03 3.69e+00 3.69e+00
1 1.07e+02 1.12e+01 1.91e+03 2.35e+00 2.18e+00
2 1.20e+02 1.93e+01 2.02e+03 1.65e+00 1.82e+00
3 1.05e+02 1.53e+01 1.80e+03 1.34e+00 1.72e+00
4 8.47e+01 4.41e+00 1.61e+03 1.16e+00 1.41e+00
5 8.02e+01 7.34e+00 1.46e+03 9.95e-01 1.29e+00
6 7.32e+01 3.71e+00 1.39e+03 8.83e-01 1.15e+00
7 7.07e+01 5.05e+00 1.31e+03 8.02e-01 1.05e+00
8 6.73e+01 4.22e+00 1.26e+03 7.36e-01 9.73e-01
9 6.53e+01 4.91e+00 1.21e+03 6.80e-01 9.01e-01
10 6.25e+01 4.65e+00 1.16e+03 6.33e-01 8.47e-01
11 6.04e+01 4.82e+00 1.11e+03 5.92e-01 7.94e-01
12 5.82e+01 4.63e+00 1.07e+03 5.55e-01 7.45e-01
13 5.62e+01 4.67e+00 1.03e+03 5.23e-01 7.04e-01
14 5.44e+01 4.59e+00 9.97e+02 4.95e-01 6.69e-01
15 5.28e+01 4.60e+00 9.64e+02 4.69e-01 6.34e-01
16 5.14e+01 4.57e+00 9.37e+02 4.46e-01 6.04e-01
17 5.01e+01 4.58e+00 9.11e+02 4.25e-01 5.75e-01
18 4.89e+01 4.57e+00 8.87e+02 4.06e-01 5.51e-01
19 4.79e+01 4.57e+00 8.66e+02 3.88e-01 5.27e-01
20 4.69e+01 4.57e+00 8.47e+02 3.72e-01 5.06e-01
21 4.61e+01 4.59e+00 8.30e+02 3.57e-01 4.86e-01
22 4.53e+01 4.60e+00 8.14e+02 3.44e-01 4.67e-01
23 4.45e+01 4.61e+00 7.98e+02 3.31e-01 4.49e-01
24 4.38e+01 4.61e+00 7.83e+02 3.19e-01 4.33e-01
25 4.31e+01 4.60e+00 7.70e+02 3.07e-01 4.17e-01
26 4.25e+01 4.60e+00 7.57e+02 2.97e-01 4.03e-01
27 4.18e+01 4.58e+00 7.45e+02 2.86e-01 3.89e-01
28 4.11e+01 4.57e+00 7.31e+02 2.77e-01 3.76e-01
29 4.05e+01 4.56e+00 7.18e+02 2.68e-01 3.65e-01
30 3.99e+01 4.54e+00 7.08e+02 2.60e-01 3.55e-01
31 3.95e+01 4.53e+00 6.99e+02 2.51e-01 3.43e-01
32 3.90e+01 4.51e+00 6.89e+02 2.44e-01 3.30e-01
33 3.84e+01 4.50e+00 6.77e+02 2.37e-01 3.20e-01
34 3.78e+01 4.50e+00 6.67e+02 2.30e-01 3.13e-01
35 3.74e+01 4.49e+00 6.59e+02 2.23e-01 3.05e-01
36 3.71e+01 4.48e+00 6.52e+02 2.17e-01 2.95e-01
37 3.67e+01 4.48e+00 6.44e+02 2.11e-01 2.85e-01
38 3.63e+01 4.47e+00 6.36e+02 2.05e-01 2.77e-01
39 3.58e+01 4.47e+00 6.28e+02 1.99e-01 2.71e-01
40 3.55e+01 4.47e+00 6.20e+02 1.94e-01 2.65e-01
41 3.51e+01 4.47e+00 6.14e+02 1.89e-01 2.58e-01
42 3.49e+01 4.47e+00 6.08e+02 1.84e-01 2.51e-01
43 3.46e+01 4.47e+00 6.02e+02 1.79e-01 2.44e-01
44 3.42e+01 4.47e+00 5.95e+02 1.75e-01 2.38e-01
45 3.39e+01 4.47e+00 5.89e+02 1.71e-01 2.33e-01
46 3.37e+01 4.46e+00 5.84e+02 1.66e-01 2.27e-01
47 3.34e+01 4.46e+00 5.79e+02 1.62e-01 2.22e-01
48 3.31e+01 4.46e+00 5.73e+02 1.58e-01 2.16e-01
49 3.28e+01 4.45e+00 5.68e+02 1.55e-01 2.11e-01
50 3.26e+01 4.45e+00 5.63e+02 1.51e-01 2.07e-01
51 3.23e+01 4.45e+00 5.58e+02 1.48e-01 2.02e-01
52 3.21e+01 4.45e+00 5.54e+02 1.44e-01 1.97e-01
53 3.19e+01 4.44e+00 5.49e+02 1.41e-01 1.93e-01
54 3.17e+01 4.44e+00 5.44e+02 1.38e-01 1.89e-01
55 3.14e+01 4.44e+00 5.40e+02 1.35e-01 1.85e-01
56 3.12e+01 4.44e+00 5.36e+02 1.32e-01 1.81e-01
57 3.11e+01 4.44e+00 5.32e+02 1.29e-01 1.77e-01
58 3.09e+01 4.44e+00 5.29e+02 1.26e-01 1.73e-01
59 3.07e+01 4.44e+00 5.25e+02 1.24e-01 1.69e-01
60 3.05e+01 4.43e+00 5.21e+02 1.21e-01 1.66e-01
61 3.03e+01 4.43e+00 5.17e+02 1.19e-01 1.63e-01
62 3.01e+01 4.43e+00 5.14e+02 1.16e-01 1.60e-01
63 3.00e+01 4.43e+00 5.11e+02 1.14e-01 1.56e-01
64 2.98e+01 4.43e+00 5.08e+02 1.11e-01 1.53e-01
65 2.96e+01 4.43e+00 5.04e+02 1.09e-01 1.50e-01
66 2.95e+01 4.43e+00 5.01e+02 1.07e-01 1.48e-01
67 2.93e+01 4.42e+00 4.98e+02 1.05e-01 1.45e-01
68 2.92e+01 4.42e+00 4.95e+02 1.03e-01 1.42e-01
69 2.90e+01 4.42e+00 4.93e+02 1.01e-01 1.39e-01
70 2.89e+01 4.42e+00 4.90e+02 9.89e-02 1.37e-01
71 2.88e+01 4.41e+00 4.87e+02 9.70e-02 1.34e-01
72 2.86e+01 4.41e+00 4.84e+02 9.52e-02 1.32e-01
73 2.85e+01 4.41e+00 4.82e+02 9.34e-02 1.29e-01
74 2.84e+01 4.41e+00 4.79e+02 9.17e-02 1.27e-01
75 2.82e+01 4.41e+00 4.77e+02 9.00e-02 1.25e-01
76 2.81e+01 4.41e+00 4.74e+02 8.84e-02 1.23e-01
77 2.80e+01 4.41e+00 4.72e+02 8.68e-02 1.21e-01
78 2.79e+01 4.41e+00 4.70e+02 8.52e-02 1.18e-01
79 2.78e+01 4.41e+00 4.67e+02 8.37e-02 1.16e-01
80 2.77e+01 4.41e+00 4.65e+02 8.22e-02 1.14e-01
81 2.76e+01 4.41e+00 4.63e+02 8.07e-02 1.13e-01
82 2.75e+01 4.40e+00 4.61e+02 7.93e-02 1.11e-01
83 2.74e+01 4.40e+00 4.59e+02 7.80e-02 1.09e-01
84 2.73e+01 4.40e+00 4.57e+02 7.66e-02 1.07e-01
85 2.72e+01 4.40e+00 4.56e+02 7.53e-02 1.05e-01
86 2.71e+01 4.40e+00 4.54e+02 7.40e-02 1.04e-01
87 2.70e+01 4.40e+00 4.52e+02 7.28e-02 1.02e-01
88 2.69e+01 4.40e+00 4.50e+02 7.16e-02 1.00e-01
89 2.68e+01 4.40e+00 4.48e+02 7.04e-02 9.89e-02
90 2.67e+01 4.40e+00 4.47e+02 6.92e-02 9.74e-02
91 2.67e+01 4.40e+00 4.45e+02 6.81e-02 9.58e-02
92 2.66e+01 4.40e+00 4.44e+02 6.70e-02 9.42e-02
93 2.65e+01 4.40e+00 4.42e+02 6.59e-02 9.27e-02
94 2.64e+01 4.40e+00 4.40e+02 6.48e-02 9.15e-02
95 2.63e+01 4.40e+00 4.39e+02 6.38e-02 9.03e-02
96 2.63e+01 4.40e+00 4.37e+02 6.28e-02 8.91e-02
97 2.62e+01 4.40e+00 4.36e+02 6.18e-02 8.77e-02
98 2.61e+01 4.39e+00 4.35e+02 6.09e-02 8.62e-02
99 2.61e+01 4.39e+00 4.33e+02 5.99e-02 8.49e-02
100 2.60e+01 4.39e+00 4.32e+02 5.90e-02 8.38e-02
101 2.59e+01 4.39e+00 4.31e+02 5.81e-02 8.27e-02
102 2.59e+01 4.39e+00 4.29e+02 5.72e-02 8.16e-02
103 2.58e+01 4.39e+00 4.28e+02 5.64e-02 8.04e-02
104 2.57e+01 4.39e+00 4.27e+02 5.55e-02 7.93e-02
105 2.57e+01 4.39e+00 4.26e+02 5.47e-02 7.83e-02
106 2.56e+01 4.39e+00 4.25e+02 5.39e-02 7.71e-02
107 2.56e+01 4.39e+00 4.23e+02 5.31e-02 7.60e-02
108 2.55e+01 4.39e+00 4.22e+02 5.23e-02 7.49e-02
109 2.54e+01 4.39e+00 4.21e+02 5.16e-02 7.40e-02
110 2.54e+01 4.39e+00 4.20e+02 5.08e-02 7.31e-02
111 2.53e+01 4.39e+00 4.19e+02 5.01e-02 7.21e-02
112 2.53e+01 4.39e+00 4.18e+02 4.94e-02 7.11e-02
113 2.52e+01 4.39e+00 4.17e+02 4.87e-02 7.01e-02
114 2.52e+01 4.39e+00 4.16e+02 4.80e-02 6.92e-02
115 2.51e+01 4.39e+00 4.15e+02 4.73e-02 6.84e-02
116 2.51e+01 4.39e+00 4.14e+02 4.66e-02 6.75e-02
117 2.50e+01 4.39e+00 4.13e+02 4.60e-02 6.65e-02
118 2.50e+01 4.39e+00 4.12e+02 4.53e-02 6.56e-02
------------------------------------------------------
Report runtimes of different methods of solving the same problem.
print("ConvBPDN solve time: %.2fs" % b.timer.elapsed('solve_wo_rsdl'))
print("ParConvBPDN solve time: %.2fs" % b_par.timer.elapsed('solve_wo_rsdl'))
print("ParConvBPDN was %.2f times faster than ConvBPDN\n" %
(b.timer.elapsed('solve_wo_rsdl')/b_par.timer.elapsed('solve_wo_rsdl')))
ConvBPDN solve time: 63.50s
ParConvBPDN solve time: 13.31s
ParConvBPDN was 4.77 times faster than ConvBPDN
Reconstruct images from sparse representations.
shr = b.reconstruct().squeeze()
imgr = sl + shr
shr_par = b_par.reconstruct().squeeze()
imgr_par = sl + shr_par
Report performances of different methods of solving the same problem.
print("Serial reconstruction PSNR: %.2fdB" % sm.psnr(img, imgr))
print("Parallel reconstruction PSNR: %.2fdB\n" % sm.psnr(img, imgr_par))
Serial reconstruction PSNR: 38.07dB
Parallel reconstruction PSNR: 38.22dB
Display original and reconstructed images.
fig = plot.figure(figsize=(21, 7))
plot.subplot(1, 3, 1)
plot.imview(img, title='Original', fig=fig)
plot.subplot(1, 3, 2)
plot.imview(imgr, title=('Serial Reconstruction PSNR: %5.2f dB' %
sm.psnr(img, imgr)), fig=fig)
plot.subplot(1, 3, 3)
plot.imview(imgr_par, title=('Parallel Reconstruction PSNR: %5.2f dB' %
sm.psnr(img, imgr_par)), fig=fig)
fig.show()
Display low pass component and sum of absolute values of coefficient maps of highpass component.
fig = plot.figure(figsize=(21, 7))
plot.subplot(1, 3, 1)
plot.imview(sl, title='Lowpass component', fig=fig)
plot.subplot(1, 3, 2)
plot.imview(np.sum(abs(X), axis=b.cri.axisM).squeeze(),
cmap=plot.cm.Blues, title='Serial Sparse Representation',
fig=fig)
plot.subplot(1, 3, 3)
plot.imview(np.sum(abs(X_par), axis=b.cri.axisM).squeeze(),
cmap=plot.cm.Blues, title='Parallel Sparse Representation',
fig=fig)
fig.show()
