You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: README.md
+15-2
Original file line number
Diff line number
Diff line change
@@ -10,8 +10,21 @@ If you used the code please cite our paper [1].
10
10
An excerpt from our paper:
11
11
12
12
13
-
"For fair comparison of the algorithms, one must take into consideration the window length of the signal, the sparsity degree of the test signal, the hyperparameters of the algorithms, such as termination criteria and desired reconstruction quality. From Fig. 4 in [1], we observed that the OMP algorithm is fastest in
14
-
reconstruction, whereas the AMP and IHT algorithms that are known to be computationally cheaper, appear to be slower. This is due to the low sparsity degree and short signal length. The OMP algorithm gives better performance for less sparse signals, and here the experiments were done with signals with less than 10% occupancy. The IHT algorithm’s performance is relatively independent from the sparsity degree, and the performance of AMP is less sensitive to sparsity degree than the OMP. These issues are rather strong practical arguments for flexible designs."
13
+
For fair comparison of the algorithms, one must take into
14
+
consideration the window length of the signal, the sparsity
15
+
degree of the test signal, the hyperparameters of the algorithms,
16
+
such as termination criteria and desired reconstruction quality.
17
+
From Fig. 4, we observed that the OMP algorithm is fastest in
18
+
reconstruction, whereas the AMP and IHT algorithms that are
19
+
known to be computationally cheaper, appear to be slower. This
20
+
is due to the low sparsity degree and short signal length. The
21
+
OMP algorithm gives better performance for less sparse signals
22
+
[27], and here the experiments were done with signals with less
23
+
than 10% occupancy. The IHT algorithm’s performance is
24
+
relatively independent from the sparsity degree, and the
25
+
performance of AMP is less sensitive to sparsity degree than the
26
+
OMP [27], [28]. These issues are rather strong practical
0 commit comments