CN103517079A - Compressed video acquisition and reconstruction system based on data driven subspace set - Google Patents
Compressed video acquisition and reconstruction system based on data driven subspace set Download PDFInfo
- Publication number
- CN103517079A CN103517079A CN201310422841.7A CN201310422841A CN103517079A CN 103517079 A CN103517079 A CN 103517079A CN 201310422841 A CN201310422841 A CN 201310422841A CN 103517079 A CN103517079 A CN 103517079A
- Authority
- CN
- China
- Prior art keywords
- subspace
- collection
- module
- sparse basis
- sparse
- 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.)
- Granted
Links
Images
Landscapes
- Image Analysis (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The invention provides a compressed video acquisition and reconstruction system based on a data driven subspace set. The compressed video acquisition and reconstruction system comprises a subspace set construction module, a sparse basis matrix construction module, a video signal sensing module and a reconstruction processing module, wherein a clustering method is utilized by the subspace set construction module to generate the subspace set, a linear subspace learning method is utilized by the sparse basis matrix construction module to generate a sparse basis corresponding to the subspace set, video signals are projected in a block mode through the sensing module, and obtained data are decoded and reconstructed in the reconstruction processing module at last. Compressed sampling is provided, meanwhile a distributed gradual model structure in a video sampling process agrees with the system, and reconstruction accuracy and efficiency of a sparse basis matrix special structure is also promoted. Sampling efficiency of video signals is greatly improved, reconstruction gains are obtain compared with other methods under different sampling compression ratios, and meanwhile the system has good expandability.
Description
Technical field
The present invention relates to a kind of vision signal and obtain scheme, specifically a kind of compressed video collection and reconfiguration system based on data-driven subspace collection.
Background technology
The collection of vision signal and coding (compression) are most important for application such as the storage of video and transmission.Traditional signal processing system adopts the pattern of first sampling recompression: in order intactly to preserve all information of signal, should to video, sample to be not less than the twice sample frequency of signal bandwidth; The primary signal collecting reaches the object of removing redundancy after by a series of coding techniquess, the bottleneck of correlation technique has been to spend a large amount of transducers and computational resource just in order to obtain a small amount of Signal Compression data after processing, too high to the resource requirement of sampling end.For the further collecting efficiency that improves vision signal, in sampling, added some signal processing technologies, wherein a kind of scheme is that sampling and compression are carried out simultaneously, then some algorithms by rear end are reconstructed the data after compressing.
Through the literature search of prior art is found, Ying Liu, in " Motion-Aware Decoding of Compressed-Sensed Video " literary composition that Ming Li and Dimitris A.Pados (TCSVT) deliver on periodical at " the IEEE Transactions on Circuits and Systems for Video Technology " of 2013, the transform(KLT based on Karhunen-Loeve having been proposed) reconstruct of base is applied to compressed sensing on video sampling, the method directly adopts sensing matrix to carry out compression sampling to video block at sample code end, in decoding end, use KLT base as sparse base, signal to be reconstructed, this method can improve the efficiency of video sampling effectively, and guarantee the subjective quality of the video that reconstruct obtains, but the KLT base that this method is used produces in Local Search window, for the video scene with complex texture or strenuous exercise, the KLT base that the method is used just can not carry out rarefaction representation to frame of video piece accurately and effectively, and then cause effect to reduce.These deficiencies impel us to remove on its basis to find a kind of reconstructing method more effectively, and the special construction that makes full use of vision signal piece improves the subjective and objective quality of reconstruction result.In " A Theory for Sampling Signals From a Union of Subspaces " literary composition that Yue M.Lu and Minh N.Do (TSP) deliver on periodical at " the IEEE Transactions on Signal Processing " of 2008, the signal sampling theory based on subspace collection has been proposed, this theory provided for be in subspace collection signal sampling will be satisfied uniqueness and the condition of stability, but the subspace collection that this theory is supposed is opened into by fixed base, can not provide more effective sparse property and adaptability.
Summary of the invention
The present invention is directed to the deficiencies in the prior art, a kind of compressed video collection and reconfiguration system based on data-driven subspace collection is provided, can effectively improve the subjective and objective quality of video signal collective efficiency and reconfiguration system, and can be used as a kind of general video acquisition instrument.
The present invention is achieved by the following technical solutions:
Compressed video collection and reconfiguration system based on data-driven subspace collection of the present invention, comprising: subspace collection constructing module, sparse basis array constructing module, vision signal sensing module and reconstruction processing module, wherein:
Described subspace collection constructing module, to vision signal key frame piece, utilizes clustering method generated subspace collection ，Bing Jianggai subspace collection to output to the input of sparse basis array constructing module;
Described sparse basis array constructing module receives subspace collection, the sparse basis array of utilizing linear subspaces learning method generated subspace set pair to answer, and this sparse basis array is outputed to the input of reconstruction processing module;
Described vision signal sensing module carries out projection to the non-key frame piece of vision signal with the form of piece, obtains measured value, and this measured value is outputed to the input of reconstruction processing module;
Described reconstruction processing module receives the basic matrix of described sparse basis array constructing module output and the measured value of described vision signal sensing module output, and signal is reconstructed.
Described subspace collection constructing module, realizes by carrying out piece cluster on the key frame rebuilding and generates different classes of piece group.Every class piece group is corresponding to a sub spaces, and the piece group that cluster obtains is the sparse basis array for generated subspace collection as training set.Subspace structure based on piece cluster can operate to realize to whole reconstruct key frame by sparse Subspace clustering method and block matching method.
Described sparse basis array constructing module is realized a kind of orthonormal basis being generated by linear subspaces learning method, and linear subspaces learning method respectively independent role obtains different bases in different piece groups, and then forms sparse basis array.It can adaptively express the immanent structure of high dimensional signal, and with respect to fixed base rarefaction representation vision signal more effectively, and the rarefaction representation of signal in this sparse basis array is to have block structured.
Described sensing module is a kind of digital micromirror device (DMD) of single order, and it has simulated the compressed sensing to vision signal.
Described reconstruction processing module is by a kind of protruding relaxed algorithm model realization.
The compressed sensing technology integrating based on data-driven subspace adopting in the present invention provides general solution as the collection of vision signal, especially for the vision signal with complex texture and strenuous exercise.Subspace used in the present invention collection is by adopting sparse subspace clustering and block matching method cluster to obtain in the key frame of reconstruct, take full advantage of the unique texture of frame of video piece, and the space time redundancy of the interior interframe of frame, improved efficiency and the performance of sampling.On the other hand, the important function of bringing into play in the restructuring procedure of compressed sensing in view of sparse basis array, the method that the present invention learns by linear subspaces learns separately to obtain the sparse basis array of corresponding base and then composition subspace collection to every sub spaces, can make like this frame block signal there is adaptability rarefaction representation, and this rarefaction representation has structural, and then raising sampling efficiency (reducing the required necessary hits of Accurate Reconstruction), can also accelerate convergence and the stability of protruding lax restructing algorithm, contribute to the performance of data-driven of the present invention subspace collection compressed sensing and the lifting of practicality.
Compared with prior art, the present invention has following beneficial effect:
The present invention has improved reconstruction property greatly, compare with the video compression sensor-based system that traditional use fixed base or KLT base are reconstructed, due to reconstruct of the present invention adopt be adaptive global optimum gene this in reconstruct effect, all can be enhanced; For other high dimensional signal, the present invention also can use by suitable modification, has stronger adaptability; When rebuilding due to the special tectonic of subspace collection and sparse basis array, make signal there is structural rarefaction representation, therefore the present invention can further improve sampling efficiency in the situation that do not reduce the subjective effect of video, accelerate the convergence rate of protruding lax restructing algorithm simultaneously, under different Sampling Compression rates, compare additive method and obtained reconstruct gain, also possess good extensibility simultaneously.
Accompanying drawing explanation
By reading the detailed description of non-limiting example being done with reference to the following drawings, it is more obvious that other features, objects and advantages of the present invention will become:
Fig. 1 is the structured flowchart of system one embodiment of the present invention;
Tu2Wei subspace collection constructing module fundamental diagram;
Fig. 3 is the structural rarefaction representation schematic diagram that sparse basis array constructing module produces frame of video block signal.
Embodiment
Below in conjunction with specific embodiment, the present invention is described in detail.Following examples will contribute to those skilled in the art further to understand the present invention, but not limit in any form the present invention.It should be pointed out that to those skilled in the art, without departing from the inventive concept of the premise, can also make some distortion and improvement.These all belong to protection scope of the present invention.
As shown in Figure 1, the structured flowchart of one embodiment of the invention, comprise: subspace collection constructing module, sparse basis array constructing module, vision signal sensing module and reconstruction processing module, wherein: subspace collection constructing module utilizes clustering method generated subspace collection, the sparse base that sparse basis array constructing module utilizes linear subspaces learning method generated subspace set pair to answer, sensing module compresses projection to vision signal with the form of piece, the finally decoded reconstruct in reconstruction processing module of the measured value of gained.In coding side, vision signal sensing module is sampled and is produced measured value vision signal; In decoding end, sparse basis array constructing module produces basic matrix; The output of described sparse basis array constructing module the measured value of basic matrix and described vision signal sensing module output together with enter reconstruction processing module, in reconstruction processing module, signal is reconstructed.
In the present embodiment, described subspace collection constructing module as shown in Figure 2, is done piece cluster in the key frame of rebuilding, wherein: the set of blocks X={x in key frame at view picture
1, x
2..., x
k, utilize sparse Subspace clustering method or block matching method that X is divided into t cluster X
1, X
2..., X
t, the piece in each cluster is similar and belongs to a sub spaces.X
1, X
2..., X
tcorresponding to t sub spaces S
1, S
2..., S
t, so arbitrary frame of video block signal x belongs to subspace collection U=∪ S
i.
In the present embodiment, described sparse basis array constructing module is realized a kind of orthonormal basis being generated by linear subspaces learning method, and linear subspaces learning method (as principal component analysis (PCA)) difference independent role is in different piece group X
i, i=1 ..., t obtains different base Ψ
i, i=1 ..., t, and then form sparse basis array Ψ
*=[Ψ
1, Ψ
2..., Ψ
t].This sparse basis array can adaptively be expressed the immanent structure of frame of video block signal, with respect to fixed base rarefaction representation vision signal more effectively, and the rarefaction representation c of signal in this sparse basis array
*to there is block structured, as shown in Figure 3.
In the present embodiment, described vision signal sensing module is a kind of digital micro-mirror projector equipment (DMD) of single order, and it has simulated the compressed sensing y=Φ x to vision signal, and Φ is stochastical sampling matrix.First this invention carries out compression sampling to key frame piece, and sample rate is 0.7, then non-key frame block signal is carried out to compression sampling, being chosen between 0.1 to 0.6 of sample rate, and the sampling based on frame of video piece has improved the speed of video sampling and reconstruct.
In the present embodiment, described reconstruction processing module, by a kind of protruding relaxed algorithm model realization, is specially: for key frame, find l
1the c of Norm minimum makes y=Φ Ψ c, and what obtain is a globally optimal solution, with two-dimensional dct base Ψ, is multiplied by the key frame block signal that this globally optimal solution just can obtain required reconstruct; For non-key frame, find l
2, Ithe c of Norm minimum
*make y=Φ Ψ
*c
*, what obtain is a globally optimal solution, uses Ψ
*be multiplied by the non-key frame block signal that this globally optimal solution just can obtain required reconstruct.Wherein, Φ is stochastical sampling matrix, l
2, Inorm is mixing norm,
i is the subscript of piece group in block structure, as Fig. 3.
Implementation result
Being set to of key parameter in the present embodiment: experiment derives from the YUV file of the 4:2:0 form of Football_cif.yuv(352x288 with video sequence), altogether get 250 frames.Every ten frames are a frame group, and choosing the first frame is key frame, and rear nine frames are non-key frame, and the selection of dimension of piece is 16 * 16 pixels.Because the gray-scale map of signal has been concentrated most energy, test mainly completes on gray-scale map.The present embodiment has compared and has adopted the methods of people in " Motion-Aware Decoding of Compressed-Sensed Video " paper such as the method for the compressed sensing based on data-driven subspace collection of the present invention and Ying Liu, and the method for people in " A Theory for Sampling Signals From a Union of Subspaces " paper such as Yue M.Lu.The present invention's sparse base used has been chosen PCA base, and the dimension of every sub spaces is 10, and the number of the subspace that cluster produces is 50.
First two method is compared with it, in compression ratio, is 0.2 o'clock, and the present embodiment system obtains respectively 9.2dB, the reconstruct gain of 2.7dB; In compression ratio, be 0.3 o'clock, the present embodiment system obtains respectively 11.6dB, the reconstruct gain of 2.8dB; In compression ratio, be 0.4 o'clock, the present embodiment system obtains respectively 11.4dB, the reconstruct gain of 4.2dB; In compression ratio, be 0.5 o'clock, the present embodiment system obtains respectively 10.3dB, the reconstruct gain of 6.3dB;
Experiment shows, the present embodiment system reconstructing video sequence out is obviously better than other two kinds of video sequences that method obtains on reconstruction quality.
Above specific embodiments of the invention are described.It will be appreciated that, the present invention is not limited to above-mentioned particular implementation, and those skilled in the art can make various distortion or modification within the scope of the claims, and this does not affect flesh and blood of the present invention.
Claims (7)
1. compressed video collection and the reconfiguration system based on data-driven subspace collection, is characterized in that, comprising: subspace collection constructing module, sparse basis array constructing module, vision signal sensing module and reconstruction processing module, wherein:
Described subspace collection constructing module, to vision signal key frame piece, utilizes clustering method generated subspace collection ，Bing Jianggai subspace collection to output to the input of sparse basis array constructing module;
Described sparse basis array constructing module receives subspace collection, the sparse basis array of utilizing linear subspaces learning method generated subspace set pair to answer, and this sparse basis array is outputed to the input of reconstruction processing module;
Described vision signal sensing module carries out projection to the non-key frame piece of vision signal with the form of piece, obtains measured value, and this measured value is outputed to the input of reconstruction processing module;
Described reconstruction processing module receives the basic matrix of described sparse basis array constructing module output and the measured value of described vision signal sensing module output, and signal is reconstructed.
2. compressed video collection and the reconfiguration system based on data-driven subspace collection according to claim 1, it is characterized in that, described subspace collection constructing module, realization generates different classes of piece group by carry out piece cluster on the key frame of reconstruct, every class piece group is corresponding to a sub spaces, and the piece group that cluster obtains is the sparse basis array for generated subspace collection as training set.
3. compressed video collection and the reconfiguration system based on data-driven subspace collection according to claim 2, it is characterized in that, the subspace structure based on piece cluster can operate to realize to whole reconstruct key frame by sparse Subspace clustering method and block matching method.
4. according to compressed video collection and reconfiguration system based on data-driven subspace collection described in claim 1-3 any one, it is characterized in that, described sparse basis array constructing module is realized a kind of orthonormal basis being generated by linear subspaces learning method, it can adaptively express the immanent structure of high dimensional signal, with respect to fixed base rarefaction representation vision signal more effectively, this rarefaction representation in sparse basis array is to have block structured.
5. compressed video collection and the reconfiguration system based on data-driven subspace collection according to claim 4, is characterized in that, described linear subspaces learning method respectively independent role obtains different bases in different piece groups, and then forms sparse basis array.
6. according to compressed video collection and reconfiguration system based on data-driven subspace collection described in claim 1-3 any one, it is characterized in that, described vision signal sensing module is a kind of digital micromirror device of single order, and it has simulated the compressed sensing to vision signal.
7. according to compressed video collection and reconfiguration system based on data-driven subspace collection described in claim 1-3 any one, it is characterized in that, described reconstruction processing module is passed through a kind of protruding relaxed algorithm model realization, and it is exactly the reconstruction signal that will obtain that the globally optimal solution finding is multiplied by sparse base.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310422841.7A CN103517079B (en) | 2013-09-16 | 2013-09-16 | Compression video acquisition based on data-driven subspace collection and reconfiguration system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310422841.7A CN103517079B (en) | 2013-09-16 | 2013-09-16 | Compression video acquisition based on data-driven subspace collection and reconfiguration system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103517079A true CN103517079A (en) | 2014-01-15 |
| CN103517079B CN103517079B (en) | 2016-11-16 |
Family
ID=49898992
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310422841.7A Active CN103517079B (en) | 2013-09-16 | 2013-09-16 | Compression video acquisition based on data-driven subspace collection and reconfiguration system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103517079B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104243986A (en) * | 2014-09-15 | 2014-12-24 | 上海交通大学 | Compressed video capture and reconstruction system based on data drive tensor subspace |
| CN104301728A (en) * | 2014-10-15 | 2015-01-21 | 上海交通大学 | Compressed video capture and reconstruction system based on structured sparse dictionary learning |
| CN105721869A (en) * | 2016-01-26 | 2016-06-29 | 上海交通大学 | Structured sparsity-based compression tensor acquisition and reconstruction system |
| US10057383B2 (en) | 2015-01-21 | 2018-08-21 | Microsoft Technology Licensing, Llc | Sparsity estimation for data transmission |
| CN110620927A (en) * | 2019-09-03 | 2019-12-27 | 上海交通大学 | Scalable compression video tensor signal acquisition and reconstruction system based on structured sparsity |
| CN110719473A (en) * | 2019-09-03 | 2020-01-21 | 上海交通大学 | Scalable compression video acquisition and reconstruction system based on structured sparsity |
| CN110944373A (en) * | 2019-09-27 | 2020-03-31 | 国家电网有限公司 | Wireless sensor network system, data transmission method, storage medium and terminal |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103237204A (en) * | 2013-03-21 | 2013-08-07 | 上海交通大学 | Video signal collection and reconfiguration system based on high-dimension compressed sensing |
-
2013
- 2013-09-16 CN CN201310422841.7A patent/CN103517079B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103237204A (en) * | 2013-03-21 | 2013-08-07 | 上海交通大学 | Video signal collection and reconfiguration system based on high-dimension compressed sensing |
Non-Patent Citations (3)
| Title |
|---|
| J. PRADES-NEBOT, Y. MA, AND T. HUANG: "Distributed Video Coding Using Compressive Sampling", 《PROCEEDINGS OF PICTURE CODING SYMPOSIUM》, 8 May 2009 (2009-05-08), pages 1 - 4 * |
| YONINA C. ELDAR AND MOSHE MISHALI: "Robust Recovery of Signals From a Structured Union of Subspaces", 《IEEE TRANSACTIONS ON INFORMATION THEORY》, 30 November 2009 (2009-11-30), pages 1 * |
| 叶鑫炜: "基于高维压缩感知的视频信号采集与重构技术研究", 《中国优秀硕士学位论文全文数据库》, 15 July 2013 (2013-07-15), pages 9 - 17 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104243986A (en) * | 2014-09-15 | 2014-12-24 | 上海交通大学 | Compressed video capture and reconstruction system based on data drive tensor subspace |
| CN104243986B (en) * | 2014-09-15 | 2018-07-20 | 上海交通大学 | Compression video acquisition and reconfiguration system based on data-driven tensor subspace |
| CN104301728A (en) * | 2014-10-15 | 2015-01-21 | 上海交通大学 | Compressed video capture and reconstruction system based on structured sparse dictionary learning |
| US10057383B2 (en) | 2015-01-21 | 2018-08-21 | Microsoft Technology Licensing, Llc | Sparsity estimation for data transmission |
| CN105721869A (en) * | 2016-01-26 | 2016-06-29 | 上海交通大学 | Structured sparsity-based compression tensor acquisition and reconstruction system |
| CN105721869B (en) * | 2016-01-26 | 2018-04-06 | 上海交通大学 | The collection of compression tensor and reconfiguration system based on structural sparse |
| CN110620927A (en) * | 2019-09-03 | 2019-12-27 | 上海交通大学 | Scalable compression video tensor signal acquisition and reconstruction system based on structured sparsity |
| CN110719473A (en) * | 2019-09-03 | 2020-01-21 | 上海交通大学 | Scalable compression video acquisition and reconstruction system based on structured sparsity |
| CN110719473B (en) * | 2019-09-03 | 2021-11-23 | 上海交通大学 | Scalable compression video acquisition and reconstruction system based on structured sparsity |
| CN110620927B (en) * | 2019-09-03 | 2022-05-27 | 上海交通大学 | Scalable compression video tensor signal acquisition and reconstruction system based on structured sparsity |
| CN110944373A (en) * | 2019-09-27 | 2020-03-31 | 国家电网有限公司 | Wireless sensor network system, data transmission method, storage medium and terminal |
| CN110944373B (en) * | 2019-09-27 | 2023-09-26 | 国家电网有限公司 | Wireless sensor network system, data transmission method, storage medium and terminal |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103517079B (en) | 2016-11-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103517079A (en) | Compressed video acquisition and reconstruction system based on data driven subspace set | |
| US10924755B2 (en) | Real time end-to-end learning system for a high frame rate video compressive sensing network | |
| Shi et al. | Image compressed sensing using convolutional neural network | |
| CN104301728A (en) | Compressed video capture and reconstruction system based on structured sparse dictionary learning | |
| CN112884851B (en) | Construction method of deep compressed sensing network based on expansion iteration optimization algorithm | |
| Yang et al. | Distributed sensor perception via sparse representation | |
| CN103237204B (en) | Based on video signal collective and the reconfiguration system of higher-dimension compressed sensing | |
| CN111565318A (en) | Video compression method based on sparse samples | |
| CN105721869B (en) | The collection of compression tensor and reconfiguration system based on structural sparse | |
| CN104199627A (en) | Gradable video coding system based on multi-scale online dictionary learning | |
| CN102123278A (en) | Signal source encoding method based on distributed compressive sensing technology | |
| CN102915562A (en) | Compressed sensing-based multi-view target tracking and 3D target reconstruction system and method | |
| CN102148986B (en) | Method for encoding progressive image based on adaptive block compressed sensing | |
| CN110728725B (en) | Hardware-friendly real-time system-oriented lossless texture compression method | |
| CN102568016B (en) | Compressive sensing image target reconstruction method based on visual attention | |
| CN104243986A (en) | Compressed video capture and reconstruction system based on data drive tensor subspace | |
| CN106993240A (en) | Many video summarization methods based on sparse coding | |
| CN111542006A (en) | Object identification method based on wireless radio frequency signal | |
| Liu et al. | Decoding of framewise compressed-sensed video via interframe total variation minimization | |
| Veluri et al. | Neuricam: Video super-resolution and colorization using key frames | |
| CN103780264A (en) | Method for compressed sensing based on FPGA | |
| CN106791871A (en) | A kind of motion vector modulation intelligence hides detection method | |
| Zhang et al. | Nuclear norm regularization with a low-rank constraint for matrix completion | |
| Lakshman et al. | Image recovery using sparse reconstruction based texture refinement | |
| CN110719473A (en) | Scalable compression video acquisition and reconstruction system based on structured sparsity |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |