CN111444470B - Distributed reconstruction method of two-channel critical sampling pattern filter bank - Google Patents
Distributed reconstruction method of two-channel critical sampling pattern filter bank Download PDFInfo
- Publication number
- CN111444470B CN111444470B CN202010235456.1A CN202010235456A CN111444470B CN 111444470 B CN111444470 B CN 111444470B CN 202010235456 A CN202010235456 A CN 202010235456A CN 111444470 B CN111444470 B CN 111444470B
- Authority
- CN
- China
- Prior art keywords
- representing
- iteration
- vector
- node
- channel
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 33
- 238000005070 sampling Methods 0.000 title claims abstract description 16
- 238000004458 analytical method Methods 0.000 claims abstract description 10
- 230000004927 fusion Effects 0.000 claims abstract description 8
- 239000013598 vector Substances 0.000 claims description 80
- 239000011159 matrix material Substances 0.000 claims description 57
- 150000001875 compounds Chemical class 0.000 claims description 9
- 238000010586 diagram Methods 0.000 claims description 7
- 238000005457 optimization Methods 0.000 abstract description 12
- 238000004088 simulation Methods 0.000 abstract description 8
- 238000004422 calculation algorithm Methods 0.000 abstract description 7
- 238000012935 Averaging Methods 0.000 abstract description 2
- 238000012545 processing Methods 0.000 description 8
- 238000004364 calculation method Methods 0.000 description 6
- 230000000694 effects Effects 0.000 description 3
- 238000012804 iterative process Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 230000015572 biosynthetic process Effects 0.000 description 2
- 230000001788 irregular Effects 0.000 description 2
- 238000003786 synthesis reaction Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000017105 transposition Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03H—IMPEDANCE NETWORKS, e.g. RESONANT CIRCUITS; RESONATORS
- H03H19/00—Networks using time-varying elements, e.g. N-path filters
- H03H19/002—N-path filters
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Computational Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computing Systems (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The invention discloses a distributed reconstruction method of a two-channel critical sampling image filter bank, which comprises the steps of firstly converting a total optimization problem into a series of local optimization problems by using a truncation operator, wherein the truncation operator is related to the relevant characteristics of an image; solving the local optimization problem to obtain a result, and performing fusion averaging; and then, iteratively updating the result obtained by fusion, eliminating errors and obtaining an optimal solution. Theoretical analysis and simulation results show that compared with the prior art, the method has the advantages of smaller complexity, smaller iteration times of the algorithm and shorter time consumption.
Description
Technical Field
The invention relates to the technical field of graph signal processing, in particular to a distributed reconstruction method of a two-channel critical sampling graph filter bank.
Background
With the advent of the big data era, irregular data and network data are emerging continuously, traditional signal processing is not enough to process the signals well, and researchers expand the traditional signal processing technology to the graph to form graph signal processing in order to process signals of irregular domains better. With continuous development, image signal processing has become a significant part of signal processing, and is receiving attention from more and more researchers.
The graph filter bank is widely noticed by researchers because of its multi-resolution characteristic, and research on the graph filter bank is also slowly evolving into an important branch of graph signal processing. The diagram filter bank is classically compared with a two-channel biorthogonal diagram filter bank (BiorGFB) having a structure as shown in FIG. 1The analysis filters of the graph filter bank are denoted as H 0 And H 1 The synthesis (reconstruction) filter can be denoted as G 0 And G 1 . When a picture signal x is input, the reconstructed signal is output
Can be expressed as:
subband signal z 0 And z 1 Can be expressed as:
z 0 ＝Q L H 0 x z 1 ＝Q H H 1 x (3)
for a bipartite graph G = (L, H, E), let i 1 ,…,i |L| Expressed as indices of vertices in the L set, i 1 ,…,i |H| Expressed as indices of vertices in the H set, where | L | + | H | = N, N is the number of vertices in the graph. Thus the sampling matrix Q L And Q H It can be defined as:
the complete reconstitution conditions were:
at a given subband signal z 0 And z 1 Thereafter, reconstruction can be achieved by optimization methods, where reconstructing a signal can be equivalent to solving a least squares problem:
the global optimal solution that can solve this problem is
Wherein
At the same time, in order to satisfy the complete reconstruction condition, the synthesis filter G 0 ,G 1 The requirements are satisfied:
as can be seen from equations (9) and (11), we can resolve such a problem as x = a -1 c, solving problems of this type is mainly based on inversion, and when the number of nodes of the graph is large (that is, the dimension of a matrix to be inverted is large), the matrix inversion can greatly increase the calculation cost, and more seriously, the matrix is possibly ill, thereby greatly affecting the robustness of the matrix. It can be seen that when the scale of the graph is large and the amount of global information is huge, such computational loss is necessarily very large.
Disclosure of Invention
The invention provides a distributed reconstruction method of a two-channel critical sampling image filter bank, aiming at the problem that the reconstruction method of the existing image filter bank has high large-scale matrix inversion calculation cost.
In order to solve the problems, the invention is realized by the following technical scheme:
the distributed reconstruction method of the two-channel critical sampling image filter bank specifically comprises the following steps:
step 2, determining a truncation operator matrix D k,2r The truncated operator matrix D k,2r Is a diagonal matrix, the diagonal elements of which are:
where C (k, 2 r) represents node k and its set of all neighbor nodes within 2r of node k;
step 3, calculating a middle matrix of the two channels; wherein:
intermediate matrix B of low channel 0 Comprises the following steps:
B 0 ＝Q L H 0
in the formula, Q L Sample matrix representing low channels, H 0 Indicating low channelsAnalyzing the filter matrix;
intermediate matrix B of high channel 1 Comprises the following steps:
B 1 ＝Q H H 1
in the formula, Q H A sampling matrix, H, representing the high channel 0 An analysis filter matrix representing a high channel;
step 4, for each node k, calculating a local reconstruction incremental vector V of the kth node under the mth iteration k,2r (m) ：
In the formula, D k,2r Representing a truncated operator matrix, B 0 Intermediate matrix representing low channels, B 1 An intermediate matrix representing the high channels,
a subband signal iteration vector representing the low channel at the mth iteration,
a subband signal iteration vector representing the high channel at the mth iteration [ ·] T Representing a transpose;
represents a pseudo-inverse;
step 5, for each node k, after carrying out average fusion on corresponding elements in local reconstruction incremental vectors of all neighbor nodes within r order of the node k and the node k, taking the node k as a reconstruction incremental vector under the mth iteration
The jth element of (1)
In the formula, | C (k, r) | represents the node k and the number of all neighbor nodes within the r order of the node k; v s,2r (m) Representing a local reconstruction incremental vector of the s-th node under the mth iteration, wherein s belongs to C (k, r), and C (k, r) represents a node k and a set of all neighbor nodes within the r-order of the node k;
step 6, updating the reconstructed image signal vector; wherein the update formula of the jth element of the reconstructed image signal vector is:
in the formula (I), the compound is shown in the specification,
representing reconstructed signal vector at m +1 th iteration
The (j) th element of (a),
representing reconstructed signal vector at mth iteration
The j element of (v) r (m) (j) The jth element of the reconstructed delta vector at the mth iteration;
step 7, judgment
Whether or not:
if the conditions are satisfied, terminating the iteration, and reconstructing the signal vector under the m +1 th iteration
Output as the final reconstruction result;
otherwise, go to step 8;
step 8, updating the sub-band signal iteration vectors of the two channels; wherein
The updating formula of the sub-band signal iteration vector of the low channel is as follows:
in the formula (I), the compound is shown in the specification,
represents the subband signal iteration vector of the low channel at the (m + 1) th iteration,
the subband signal iteration vector, B, representing the low channel at the mth iteration 0 The intermediate matrix representing the low channel is,
representing a reconstructed delta vector;
the updating formula of the sub-band signal iteration vector of the high channel is as follows:
in the formula (I), the compound is shown in the specification,
represents the subband signal iteration vector of the high channel at the m +1 th iteration,
the subband signal iteration vector representing the high channel at the mth iteration, B 1 An intermediate matrix representing the high channels,
representing a reconstructed delta vector;
step 9, enabling the iteration times to be m +1, and returning to the step 4;
k represents the node number of the graph, j represents the number of the corresponding vector, j, k =1,2 \8230, and N are the number of the nodes of the graph.
In the above scheme, the node k and the set C (k, r) of all neighboring nodes within the r-th order of the node k need to satisfy the following condition:
wherein:
representing the bearling dimension of the diagram,
representing the Berling density of the graph and r representing the neighborhood radius.
In the above scheme, the node k and the set C (k, 2 r) of all neighboring nodes within 2r order of the node k need to satisfy the following condition:
wherein:
representing the bearling dimension of the diagram,
representing the Berling density of the graph, and 2r the neighborhood radius.
Compared with the prior art, the invention firstly uses a truncation operator to convert the total optimization problem into a series of local optimization problems, and the truncation operator is related to the relevant characteristics of the graph; solving the local optimization problem to obtain a result, and performing fusion averaging; and then, iterating the result obtained by fusion, eliminating errors and obtaining an optimal solution. Theoretical analysis and simulation results show that compared with the prior art, the method has the advantages of smaller complexity, smaller iteration times of the algorithm and shorter time consumption.
Drawings
Fig. 1 is a block diagram of a two-channel critical sampling pattern filter bank.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to the accompanying drawings in conjunction with specific examples.
According to at x * ＝A -1 In the step c, when the matrix dimension is too large, the matrix inversion is complex, and therefore calculation is complex, a truncation operator is provided to convert a total optimization problem into a series of small local optimization problems to solve, and the matrix dimension is reduced.
In the general optimization problem of (8), the truncation operator D is added k,2r It is a diagonal matrix whose diagonal elements are defined as:
where C (k, 2 r) represents the set of all neighbor nodes within 2r order of node k, where C (k, 2 r) needs to satisfy
Wherein
The method is characterized in that the Bergling dimension and the Bergling density of the graph are shown, when the Bergling dimension and the Bergling density are small, the number of neighbor nodes cannot be fast increased along with the increase of r, and the graph is sparse and can be operated in a distributed mode. The optimization problem in equation (8) then becomes a small set of optimization problems as follows:
wherein
Represents the square of the two-norm of a matrix or vector;
then we can find the local optimal solution V k,2r ：
Wherein
Representing the pseudo-inverse of the matrix.
Although the information of the 2r order neighbors is selected to solve the local optimal solution, when the local optimal solution is fused, the local optimal solution of the 2r order neighbors cannot be selected to be averaged, because a boundary effect occurs, that is, the signal value of the boundary node deviates from the original value to a large extent. Therefore, only the information of the r-order neighbor is selected for fusion.
Where C (k, r) denotes the number of r-th order neighbors of the k node, C (k, r) needs to satisfy
This is the value of each node k, which, in combination with equation (14), can be written in matrix form:
wherein
And updating iteration, wherein an iteration formula is as follows:
in the iterative process, the node k needs to store the information of the node k and the 2 r-order neighbor at the beginning
And then based on this information, local calculations are performed, requiring a calculated amount
Then transmitting the calculated result to its r-order neighbor and receiving the information about itself from the neighbor, which is calculated by the neighbor node; and then calculating and updating the information of the neighbor according to the information of the neighbor, comparing the information with the original information, if the difference between the previous information and the next information is less than the iteration condition epsilon, indicating that the information is the optimal solution of the information at the point, and if the difference is not more than the iteration condition epsilon, continuing to exchange information with the neighbor. The computational complexity of the whole iterative process is therefore: when r is small, and the graph is sparse (i.e., the graph is sparse)
Relatively small), the complexity of the entire iterative process is more nearly linear with respect to the graph scale N than the original inverted O (N) 3 ) Compared with the complexity of the method, the calculation amount is obviously reduced. Theoretical analysis shows that compared with the prior invention, the algorithm of the invention has smaller complexity, fewer iteration times and shorter program running time.
Based on the above analysis, the distributed reconstruction method of the two-channel critical sampling pattern filter bank designed by the invention specifically comprises the following steps:
step 2, determining a truncation operator matrix D k,2r The truncated operator matrix D k,2r Is a diagonal matrix, the diagonal elements of which are:
where C (k, 2 r) represents node k and its set of all neighbor nodes within 2r of node k; c (k, 2 r) is required to satisfy the following conditions
Step 3, calculating a middle matrix of the two channels; wherein:
intermediate matrix B of low channel 0 Comprises the following steps:
B 0 ＝Q L H 0 (20)
in the formula, Q L Sample matrix representing low channels, H 0 An analysis filter matrix representing a low channel;
high-channel intermediate matrix B 1 Comprises the following steps:
B 1 ＝Q H H 1 (21)
in the formula, Q H A sampling matrix, H, representing the high channel 0 Analytic filter moment representing high channelArraying;
step 4, for each node k, calculating a local reconstruction incremental vector V of the kth node under the mth iteration k,2r (m) ：
In the formula, D k,2r Representing a truncated operator matrix, B 0 Intermediate matrix representing low channels, B 1 An intermediate matrix representing the high channels,
the subband signal iteration vector representing the low channel at the mth iteration,
a subband signal iteration vector representing the high channel at the mth iteration [ ·] T Representing a transposition;
representing a pseudo-inverse;
step 5, for each node k, after carrying out average fusion on corresponding elements in local reconstruction incremental vectors of all neighbor nodes within r order of the node k and the node k, taking the node k as a reconstruction incremental vector under the mth iteration
The jth element of (1)
In the formula, | C (k, r) | represents the node k and the number of all neighbor nodes within the r-th order of the node k; v s , 2r (m) Represents the locally reconstructed delta vector of the s-th node at the m-th iteration, s ∈ C (k, r)) C (k, r) represents a node k and a set of all neighbor nodes within the r-th order of the node k, and the C (k, r) needs to satisfy the following condition
V s,2r (m) And V k,2r (m) Similarly, the result is obtained by equation (22).
Step 6, updating the reconstructed image signal vector; wherein the update formula of the jth element of the reconstructed picture signal vector is:
in the formula (I), the compound is shown in the specification,
represents the reconstructed signal vector at the (m + 1) th iteration
The (j) th element of (a),
representing reconstructed signal vectors at the m-th iteration
The j element of (v) r (m) (j) The jth element of the reconstructed delta vector at the mth iteration;
step 7, judging | v r (m) (j) Whether | ≦ ε holds:
if the conditions are satisfied, terminating the iteration, and reconstructing the signal vector under the m +1 th iteration
Output as the final reconstruction result;
otherwise, go to step 8;
step 8, updating the sub-band signal iteration vectors of the two channels; wherein
The updating formula of the sub-band signal iteration vector of the low channel is as follows:
in the formula (I), the compound is shown in the specification,
represents the subband signal iteration vector of the low channel at the (m + 1) th iteration,
the subband signal iteration vector representing the low channel at the mth iteration, B 0 An intermediate matrix representing the low channel is shown,
representing a reconstructed delta vector;
the updating formula of the sub-band signal iteration vector of the high channel is as follows:
in the formula (I), the compound is shown in the specification,
represents the subband signal iteration vector of the high channel at the (m + 1) th iteration,
the subband signal iteration vector representing the high channel at the mth iteration, B 1 An intermediate matrix representing the high channel,
representing a reconstructed delta vector;
and 9, enabling the iteration times to be m +1, and returning to the step 4.
The performance of the present invention is illustrated by the following specific simulation examples.
In simulation experiments of a graph filter bank, a Minnesota traffic map is usually used as a test object, and a classical gradient method, a maximum spanning tree sampling-based bi-orthogonal graph filter bank design method (MST-Bior) and a distributed reconstruction algorithm of the invention are applied to respectively complete reconstruction and denoising of a two-channel critical sampling map filter bank of the Minnesota traffic map.
(1) Reconstruction
In the simulation example, the signal on the graph is a test signal, the element is 1 or-1, the analysis filter bank uses a first-order spline filter bank, the evaluation index is the signal-to-noise ratio of input and output, and the expression is as follows:
in the simulation, table 1 shows the signal-to-noise ratios obtained with different methods for reconstruction, since the invention is also an iterative algorithm in practice, a comparison with the classical gradient method was chosen. From table 1, it can be seen that the snr obtained by the distributed reconstruction algorithm proposed by the present invention is close to that obtained by the centralized method, which indicates that we can achieve the same effect as the centralized method, and meanwhile, compared with the gradient method, the number of iterations is significantly reduced, and the simulation result indicates that the algorithm is low in complexity.
TABLE 1 reconstitution Properties obtained by different methods
| Signal to noise ratio/dB | Number of iterations | |
| Classical gradient method | 304.63 | 393 |
| MST-Bior | 305.14 | ---- |
| The invention | 305.01 | 22 |
(2) De-noising
In the simulation experiment, an input signal is a noisy signal (composed of a test signal and a gaussian noise signal), and as is known, noise belongs to a high-frequency signal for an original signal, a threshold value processing can be performed on a sub-band signal of a high-frequency channel, and then reconstruction is performed, so that the noise can be removed. Table 2 shows the comparison of the denoising performance of different methods at different standard deviations. As can be seen from Table 2, all the methods can perform denoising well, and at a medium noise level, the denoising effect of the existing method is better than that of the present invention, but the rest is better. Compared with the prior art, the method has the advantages that the calculation complexity is lower, the method is more suitable for the operation of large-scale graphs, the time consumption is lower, and meanwhile, the accuracy can be ensured.
TABLE 2 De-noising Performance obtained by different methods under the Minnesota Chart
It should be noted that, although the above-mentioned embodiments of the present invention are illustrative, the present invention is not limited thereto, and therefore, the present invention is not limited to the above-mentioned specific embodiments. Other embodiments, which can be made by those skilled in the art in light of the teachings of the present invention, are considered to be within the scope of the present invention without departing from its principles.
Claims (3)
1. The distributed reconstruction method of the two-channel critical sampling pattern filter bank is characterized by comprising the following steps of:
step 1, initialization: make initial low-channel sub-band signal iteration vector
Wherein z is 0 A subband signal observation vector representing a low channel; iterative vector of initial high-channel sub-band signals
Wherein z is 1 Representing a sub-band signal observation vector of a high channel; let the original reconstructed image signal vector
The jth element of (1)
Let j = k; let initial number of iterations m =0; simultaneously giving an iteration termination condition epsilon;
step 2, determining a truncation operator matrix D k,2r The truncated operator matrix D k,2r Is a diagonal matrix, the diagonal elements of which are:
where C (k, 2 r) represents node k and its set of all neighbor nodes within 2r of node k;
step 3, calculating a middle matrix of the two channels; wherein:
intermediate matrix B of low channel 0 Comprises the following steps:
B 0 ＝Q L H 0
in the formula, Q L Representing a low passSampling matrix of tracks, H 0 An analysis filter matrix representing a low channel;
high-channel intermediate matrix B 1 Comprises the following steps:
B 1 ＝Q H H 1
in the formula, Q H A sampling matrix, H, representing the high channel 1 An analysis filter matrix representing a high channel;
step 4, for each node k, calculating a local reconstruction incremental vector V of the kth node under the mth iteration k,2r (m) ：
In the formula D k,2r Representing a truncated operator matrix, B 0 Intermediate matrix representing the low channel, B 1 An intermediate matrix representing the high channels,
the subband signal iteration vector representing the low channel at the mth iteration,
a subband signal iteration vector representing the high channel at the mth iteration, [. Cndot] T Representing a transpose;
represents a pseudo-inverse;
step 5, for each node k, after carrying out average fusion on corresponding elements in local reconstruction incremental vectors of all neighbor nodes within r order of the node k and the node k, taking the node k as a reconstruction incremental vector under the mth iteration
The j (th) element of (1)
In the formula, | C (k, r) | represents the node k and the number of all neighbor nodes within the r order of the node k; v s,2r (m) Representing the local reconstruction increment vector of the s-th node under the mth iteration, wherein s belongs to C (k, r), and C (k, r) represents the node k and the set of all neighbor nodes within the r-th order of the node k;
step 6, updating the reconstructed image signal vector; wherein the update formula of the jth element of the reconstructed picture signal vector is:
in the formula (I), the compound is shown in the specification,
representing reconstructed signal vector at m +1 th iteration
The (j) th element of (a),
representing reconstructed signal vectors at the m-th iteration
The j element of (v) r (m) (j) The jth element of the reconstructed delta vector at the mth iteration;
step 7, judging | v r (m) (j) Whether | ≦ ε holds:
if the conditions are satisfied, terminating the iteration, and reconstructing the signal vector under the m +1 th iteration
Output as the final reconstruction result;
otherwise, go to step 8;
step 8, updating the sub-band signal iteration vectors of the two channels; wherein:
the updating formula of the sub-band signal iteration vector of the low channel is as follows:
in the formula (I), the compound is shown in the specification,
represents the subband signal iteration vector of the low channel at the (m + 1) th iteration,
the subband signal iteration vector representing the low channel at the mth iteration, B 0 An intermediate matrix representing the low channel is shown,
representing a reconstructed delta vector;
the update formula of the sub-band signal iteration vector of the high channel is as follows:
in the formula (I), the compound is shown in the specification,
represents the subband signal iteration vector of the high channel at the m +1 th iteration,
a subband signal iteration vector representing the high path at the mth iteration, B 1 An intermediate matrix representing the high channels,
representing a reconstructed delta vector;
9, enabling the iteration times to be m +1, and returning to the step 4;
k represents a node number of the graph, j represents a vector number, j, k =1,2 \8230, and N, N represents the number of nodes of the graph.
2. The method of claim 1, wherein the set C (k, r) of node k and all neighboring nodes within r order of node k is required to satisfy the following condition:
wherein:
representing the bearling dimension of the diagram,
representing the Berling density of the graph and r representing the neighborhood radius.
3. The distributed reconstruction method of the two-channel critical sampling pattern filter bank according to claim 1, wherein the set C (k, 2 r) of node k and all neighboring nodes within 2r order of node k is required to satisfy the following condition:
wherein:
representing the bearling dimension of the diagram,
representing the Berling density of the graph, and 2r the neighborhood radius.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010235456.1A CN111444470B (en) | 2020-03-30 | 2020-03-30 | Distributed reconstruction method of two-channel critical sampling pattern filter bank |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010235456.1A CN111444470B (en) | 2020-03-30 | 2020-03-30 | Distributed reconstruction method of two-channel critical sampling pattern filter bank |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111444470A CN111444470A (en) | 2020-07-24 |
| CN111444470B true CN111444470B (en) | 2023-03-14 |
Family
ID=71649326
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010235456.1A Active CN111444470B (en) | 2020-03-30 | 2020-03-30 | Distributed reconstruction method of two-channel critical sampling pattern filter bank |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111444470B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112560243B (en) * | 2020-12-07 | 2022-11-15 | 桂林电子科技大学 | Design method for improving frequency domain critical sampling image filter bank |
| CN112818526B (en) * | 2021-01-20 | 2022-09-30 | 桂林电子科技大学 | Distributed design method of non-uniform graph filter bank |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107256537A (en) * | 2017-06-06 | 2017-10-17 | 桂林电子科技大学 | A kind of design method for designing two passage orthogonal graph wave filter groups |
| WO2018068629A1 (en) * | 2016-10-10 | 2018-04-19 | 深圳超级数据链技术有限公司 | Signal sampling and recovery method and apparatus applicable to ovxdm system and ovxdm system |
| CN109035165A (en) * | 2018-07-16 | 2018-12-18 | 桂林电子科技大学 | The distributed figure signal reconfiguring method shunk based on neighborhood |
| CN109586688A (en) * | 2018-12-07 | 2019-04-05 | 桂林电子科技大学 | Time-varying based on iterative calculation can divide the design method of non-lower sampling figure filter group |
| CN110807255A (en) * | 2019-10-30 | 2020-02-18 | 桂林电子科技大学 | Optimization design method of M-channel joint time vertex non-downsampling filter bank |
-
2020
- 2020-03-30 CN CN202010235456.1A patent/CN111444470B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018068629A1 (en) * | 2016-10-10 | 2018-04-19 | 深圳超级数据链技术有限公司 | Signal sampling and recovery method and apparatus applicable to ovxdm system and ovxdm system |
| CN107256537A (en) * | 2017-06-06 | 2017-10-17 | 桂林电子科技大学 | A kind of design method for designing two passage orthogonal graph wave filter groups |
| CN109035165A (en) * | 2018-07-16 | 2018-12-18 | 桂林电子科技大学 | The distributed figure signal reconfiguring method shunk based on neighborhood |
| CN109586688A (en) * | 2018-12-07 | 2019-04-05 | 桂林电子科技大学 | Time-varying based on iterative calculation can divide the design method of non-lower sampling figure filter group |
| CN110807255A (en) * | 2019-10-30 | 2020-02-18 | 桂林电子科技大学 | Optimization design method of M-channel joint time vertex non-downsampling filter bank |
Non-Patent Citations (3)
| Title |
|---|
| Iterative Chebyshev Polynomial Algorithm for Signal Denoising on Graphs;Cheng Cheng等;《 2019 13th International conference on Sampling Theory and Applications (SampTA)》;20200312;全文 * |
| 一种设计 M 通道双正交过采样图滤波器组的新算法;蒋俊正等;《电子与信息学报》;20171231;全文 * |
| 两通道正交图滤波器组设计新算法;蒋俊正等;《西安电子科技大学学报》;20170928(第02期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111444470A (en) | 2020-07-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111444470B (en) | Distributed reconstruction method of two-channel critical sampling pattern filter bank | |
| Chow et al. | Ck centre unstable manifolds | |
| CN110082708B (en) | Non-uniform array design and direction of arrival estimation method | |
| CN109586688B (en) | Design method of time-varying separable non-downsampling image filter bank based on iterative computation | |
| CN107256537B (en) | Design method for designing two-channel orthogonal diagram filter bank | |
| CN113572708B (en) | DFT channel estimation improvement method | |
| CN107918710B (en) | Convex optimization-based design method of non-downsampling image filter bank | |
| CN101697150A (en) | Lifting scheme-based optimization implementation method of 9/7 wavelet transformation | |
| CN113222860B (en) | Image recovery method and system based on noise structure multiple regularization | |
| CN113222812A (en) | Image reconstruction method based on information flow reinforced deep expansion network | |
| CN109636722A (en) | A method of the online dictionary learning super-resolution rebuilding based on rarefaction representation | |
| CN102915735A (en) | Noise-containing speech signal reconstruction method and noise-containing speech signal device based on compressed sensing | |
| Zeng et al. | Bipartite subgraph decomposition for critically sampled wavelet filterbanks on arbitrary graphs | |
| CN112818526B (en) | Distributed design method of non-uniform graph filter bank | |
| CN111929641A (en) | Rapid indoor fingerprint positioning method based on width learning | |
| CN108711143A (en) | The design method of the two channel non-lower sampling figure filter groups based on relatively prime polynomials | |
| CN109856673B (en) | High-resolution Radon transformation data separation technology based on dominant frequency iterative weighting | |
| CN116186498A (en) | Earthquake signal noise suppression method and system based on self-supervision learning | |
| CN115935888A (en) | Neural network accelerating system | |
| CN113609809B (en) | Method, system, equipment, medium and terminal for diagnosing faults of radio frequency low-noise discharge circuit | |
| CN113344245A (en) | Hybrid deep learning short-term prediction model, method, storage medium, and computing device | |
| CN113191978B (en) | Asynchronous implementation method of distributed constraint edge-varying FIR (finite impulse response) graph filter | |
| CN114584150B (en) | Sensor network signal complement method realized by deep neural network | |
| Yang et al. | Graph signal reconstruction based on spatio-temporal features learning | |
| CN115327624B (en) | Inversion method and inversion system for seismic wavelets and reflection coefficients |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| EE01 | Entry into force of recordation of patent licensing contract |
Application publication date: 20200724 Assignee: Guangxi wisdom Valley Technology Co.,Ltd. Assignor: GUILIN University OF ELECTRONIC TECHNOLOGY Contract record no.: X2023980046615 Denomination of invention: A Distributed Reconstruction Method for Two Channel Critical Sampling Graph Filter Banks Granted publication date: 20230314 License type: Common License Record date: 20231108 |
|
| EE01 | Entry into force of recordation of patent licensing contract |