CN107273908A - A kind of target identification method based on compressed sensing - Google Patents
A kind of target identification method based on compressed sensing Download PDFInfo
- Publication number
- CN107273908A CN107273908A CN201710548175.XA CN201710548175A CN107273908A CN 107273908 A CN107273908 A CN 107273908A CN 201710548175 A CN201710548175 A CN 201710548175A CN 107273908 A CN107273908 A CN 107273908A
- Authority
- CN
- China
- Prior art keywords
- target
- matrix
- dictionary
- kinds
- targets
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
- G06V10/457—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components by analysing connectivity, e.g. edge linking, connected component analysis or slices
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/24—Character recognition characterised by the processing or recognition method
- G06V30/242—Division of the character sequences into groups prior to recognition; Selection of dictionaries
Abstract
The invention discloses a kind of target identification method based on compressed sensing, including step：Obtain at least two classification target master sample figures；All kinds of clarification of objective atoms are obtained using characteristic atomic extracting method；Each characteristic atomic of target described in every class is diagonally rearranged into every classification target dictionary Ψ respectivelyp, and the dictionary arranged in parallel of all kinds of targets is constituted to comprehensive dictionary Ψ；Sampling is compressed to original image x to be identified using calculation matrix Φ, the sampled signal y compressed；With reference to comprehensive dictionary Ψ, calculation matrix Φ and sampled signal y, the sparse coefficient θ of original image to be identified is obtained by reconstruction calculations；Sparse coefficient θ progress is handled and obtains coefficient figure, the line number and size according to residing for connected domain in coefficient figure carry out Classification and Identification and counting, to realize the identification to all kinds of targets in original image.Target identification method proposed by the present invention, it is possible to increase the speed of target identification in the picture, it is possible to realize multi-targets recognition.
Description
Technical field
The present invention relates to image identification technical field, more particularly to a kind of target identification method based on compressed sensing.
Background technology
It is by specific target is distinguished in image mistake using various programmed algorithms to carry out target identification in the picture
Journey, and using the target distinguished as offer basis is further processed, can be extensive in today of informatization and network
It is applied to many fields.Human eye speed when some specific objective is identified is often relatively slow, and is carried out for similar target
Identification is divided for a long time, be will also result in aestheticly tired and is gradually produced a large amount of wrong identifications, and uses machine recognition to replace people's outlook
Not, the use brain volume of human eye is replaced to improve speed and reduction energy consumption using computer amount of calculation, for field of image recognition
Speech be it is very favorable, for example：The frame of video picture of 1,000 width crossroads is identified, it is desirable to find out the wagon flow passed through
Amount, hence it is evident that eye recognition is much conducive to using machine recognition；If likewise, to robot add images steganalysis system,
It is also very favorable for development artificial intelligence technology then equivalent to the addition of " eyes " to robot.Current this area skill
Art personnel are made that many contributions in terms of target identification, and image recognition technology is not only applied to recognition of face, article by people
In terms of identification, also applied in terms of handwriting recognition, be very easy to the life of people.
And images steganalysis technology conventional at present typically has the disadvantage that time-consuming longer, speed is slower, reason is traditional figure
As target identification technology needs below scheme：Image preprocessing, image segmentation, feature extraction and feature recognition or matching；Need
The amount of image information to be obtained to sampling, which carries out repeatedly processing, to come out required target identification, and existing image is known
Image acquisition process is not separated with target identification process, is unfavorable for the raising of speed.
The disclosure of background above technology contents is only used for design and the technical scheme that auxiliary understands the present invention, and it is not necessarily
Belong to the prior art of present patent application, without tangible proof show the above present patent application the applying date
In the case of disclosed, above-mentioned background technology should not be taken to evaluate the novelty and creativeness of the application.
The content of the invention
In order to solve the above technical problems, the present invention proposes a kind of target identification method based on compressed sensing, it is possible to increase
The speed of target identification in the picture, it is possible to realize multi-targets recognition.
In order to achieve the above object, the present invention uses following technical scheme：
The invention discloses a kind of target identification method based on compressed sensing, comprise the following steps：
S1：Obtain at least two classification target master sample figures；
S2：According to the master sample figure of all kinds of targets, all kinds of targets are obtained using characteristic atomic extracting method
Characteristic atomic；
S3：Each described characteristic atomic of target described in every class is diagonally rearranged into the dictionary of target described in every class respectively
Ψp, and the dictionary arranged in parallel of all kinds of targets is constituted to comprehensive dictionary Ψ；
S4：Sampling is compressed to original image x to be identified using calculation matrix Φ, the sampled signal y compressed；
S5：With reference to comprehensive dictionary Ψ, calculation matrix Φ and sampled signal y, institute to be identified is obtained by reconstruction calculations
State the sparse coefficient θ of original image；
S6：Sparse coefficient θ progress is handled and obtains coefficient figure, line number according to residing for connected domain in the coefficient figure and
Size carries out Classification and Identification and counting, to realize the identification to all kinds of targets in the original image.
Preferably, step S1 is specifically included：Target described at least two classes existed according to known in image or video flowing, is carried
The target image of all kinds of targets is taken out, multiple target images of target described in every class are weighted respectively
To the master sample figure of target described in every class.
Preferably, wherein the target image for extracting all kinds of targets is specially：Using morphological image method or people
Work recognizes dividing method to extract the target image of all kinds of targets.
Preferably, step S4 also includes sampled signal y being stored as data in memory.
Preferably, step S5 is specifically included：Obtained perceiving matrix A according to comprehensive dictionary Ψ and calculation matrix Φ product
=Φ * Ψ, with reference to sampled signal y, are calculated by the reconstruction calculations formula y=A* θ of OMP algorithms and obtain the dilute of the original image
Sparse coefficient θ.
Preferably, the target identification method also includes step S7：Sparse coefficient θ is multiplied with comprehensive dictionary Ψ, obtained
To the reconstructed image x of collection1=Ψ * θ.
Preferably, the characteristic atomic extracting method in step S2 is specifically extracted described using MOD algorithms or K-SVD algorithms
Clarification of objective atom.
Preferably, the characteristic atomic extracting method in step S2 is specifically included：
S21：The objective matrix m of the master sample figure of all kinds of targets is inputted, and is initialized：Matrix n is initial
Turn to n=m, circulation mark i=0, the dictionary Ψ of the target initiallypFor empty set；
S22：Assignment circulation mark i=i+1；
S23：The maximum kth row of two norms in matrix n all row are found, the column element λ of extraction is used asi, whereinS is matrix n columns；
S24：The column element λ that all row are arranged with kth in calculating matrix niOptimal ratio t, whereinJ refers to the jth row in matrix n；
S25：Update the residual error r respectively arranged in matrix n residual error r, residual error rjCalculation formula be rj=nj-λi*tj；
S26：By the residual error r of all row in matrix njWith first threshold ξ1It is compared, if there is rj＜ ξ1, perform step
S27, otherwise return to step S23；
S27：Matrix n is updated to n=r, and deletes nj, dictionary ΨpIt is updated to Ψp=[Ψp, λi]；
S28：By two norms of the matrix n after renewal and Second Threshold ξ2It is compared, if | | n | | ＜ ξ2, perform step
S29, otherwise return to step S22；
S29：By dictionary ΨpIt is orthogonalized processing, output characteristic atom.
Preferably, S210 is also included in step S2：The characteristic atomic combining target matrix m of output is subjected to OMP reconstruct meters
Calculation obtains corresponding coefficient, and by corresponding coefficient, often two norms of row calculating are sized, according to coefficient magnitude and error
Effectively, then the corresponding corresponding preceding N of characteristic atomic for judging to extract arranges defeated as final characteristic atomic for N rows before sets requirement judges
Go out.
Preferably, orthogonalization process specifically uses Smith's orthogonalization process in step S29；Preferably, ξ1≤ξ2。
Compared with prior art, the beneficial effects of the present invention are：In the target identification method of the present invention, pass through spy first
Levy atom extracting method and obtain all kinds of clarification of objective atoms, will diagonally be rearranged respectively per each characteristic atomic of classification target
Comprehensive dictionary is constituted per classification target dictionary, then by the dictionary arranged in parallel of all kinds of targets, by the way that characteristic atomic is carried out into this
The particular arrangement of sample, coefficient figure is obtained with reference to sparse coefficient, is recognized and is counted directly in coefficient data, you can point door sorts out area
How many each target of every class target is separated, so as to carry out the identifying processing of multiple targets simultaneously；In addition compared to traditional
Target identification method, Classification and Identification is carried out without extracting area-of-interest progress matching primitives, directly specific by setting
Dictionary just can obtain inhomogeneity target and be respectively at different coefficient regions, i.e., counting is identified in coefficient data, treat
Journey is simple, and amount of calculation is small, improves the speed of target identification in the picture.
In further scheme, reconstructed image can be also obtained parallel while target identification is carried out, parallel computation is mutual
Do not influence, IMAQ is united two into one with target identification process, target identification is carried out in image pick-up signal, further
The speed of target identification in the picture is improved, and improves storage efficiency and the consumption of reduction hardware of data.More entering
In the scheme of one step, the principles such as limited isometry in compressive sensing theory are combined, for required identification target and background ring
The characteristics of border ga s safety degree, devises specific characteristic atomic extracting method, and by the way that characteristic atomic is combined into shape by rational
Into dictionary, the coefficient matrix for calculating and obtaining is combined with calculation matrix with this dictionary, more preferable target identification effect can be obtained.
Brief description of the drawings
Fig. 1 is the schematic flow sheet of the target identification method based on compressed sensing of the preferred embodiment of the present invention；
Fig. 2 is the schematic flow sheet of the characteristic atomic extracting method of the preferred embodiment of the present invention.
Embodiment
Below against accompanying drawing and with reference to preferred embodiment the invention will be further described.
As shown in figure 1, the preferred embodiments of the present invention disclose a kind of target identification method based on compressed sensing, including
Following steps：
S1：Obtain at least two classification target master sample figures；
At least two class targets existed according to known in some images or video flowing, are extracted by image phychology method
Target image, or manual identified are partitioned into the target image in image；Then it will distinguish per the multiple target images of classification target
It is weighted and obtains every classification target master sample figure.
S2：According to the master sample figure of all kinds of targets, all kinds of clarifications of objective are obtained using characteristic atomic extracting method former
Son；
Wherein characteristic atomic extracting method can extract all kinds of clarifications of objective originals using MOD algorithms or K-SVD algorithms
Son, further, characteristic atomic extracting method also can be using the methods told about hereinafter.
S3：Every classification target dictionary Ψ will be diagonally rearranged respectively per each characteristic atomic of classification targetp, and will be all kinds of
The dictionary Ψ of the dictionary arranged in parallel composition synthesis of target；
S4：Sampling is compressed to original image x to be identified using calculation matrix Φ, the sampled signal y compressed,
And sampled signal y is stored as data in memory；
S5：With reference to comprehensive dictionary Ψ, calculation matrix Φ and sampled signal y, original to be identified is obtained by reconstruction calculations
The sparse coefficient θ of beginning image；
Specifically, obtained perceiving matrix A=Φ * Ψ according to comprehensive dictionary Ψ and calculation matrix Φ product, with reference to adopting
Sample signal y, the sparse coefficient θ for obtaining original image is calculated by the reconstruction calculations formula y=A* θ of OMP algorithms.
S6：Sparse coefficient θ processing is obtained into coefficient figure, the line number and size according to residing for connected domain in coefficient figure are carried out
Classification and Identification and counting, to realize the identification to all kinds of targets in original image；
Specifically, sparse coefficient θ is filtered the processing such as binaryzation and obtains coefficient figure.
In a further embodiment, also include in step S5 the sparse coefficient theta replication of original image is a；Step S6
In obtain coefficient figure for the sparse coefficient θ progress of a copy of it is handled；The target identification method also includes step S7：Will in addition
A sparse coefficient θ is multiplied with comprehensive dictionary Ψ, the reconstructed image x gathered1=Ψ * θ, wherein step S7 and step S6
It can carry out simultaneously.The reconstructed image obtained by step S7 is not lost compared with the image that traditional Nyquist methods are collected
Very, original image can be substituted and carry out relevant treatment.
In order to improve target identification speed and the data volume of reduction IMAQ, the present invention uses random Gaussian or Bernoulli Jacob
Calculation matrix carries out data compression collection, then collects compared to the storage of original image less data into memory, so
Several times more of image information can be gathered under same amount of storage；Then, by the signal collected when needed from memory
It is middle to be reconstructed with OMP algorithms come by the atom that algorithm that selected dictionary is designed, designed is generated passes through particular arrangement
Composition (each characteristic atomic is in diagonal row mode in per classification target dictionary, is then arranged side by side per classification target dictionary
The comprehensive dictionary of row composition), thus calculate obtained sparse coefficient slightly through processing just divide door to sort out to distinguish every class target each
How many target, directly mutually can obtain and original image matter at convenience while calculating obtained sparse coefficient with selected dictionary
The close reconstructed image of amount, and obtain reconstruct image and can carry out parallel processing with classification identification and be independent of each other.
In a still further embodiment, the characteristic atomic extracting method in step S2 is specifically included：
S21：The objective matrix m of the master sample figure of all kinds of targets is inputted, and is initialized：Matrix n is initialized as n
=m, circulation mark i=0, the dictionary Ψ of initial targetpFor empty set；
S22：Assignment circulation mark i=i+1, the step is the dictionary Ψ of carrying targetpHow many row contained；
S23：The maximum kth row of two norms in matrix n all row are found, the column element λ of extraction is used asi, whereinS is matrix n columns, and the step is to obtain the maximum row of influence factor in matrix n；
S24：The column element λ that all row are arranged with kth in calculating matrix niOptimal ratio tj, whereinJ refers to the jth row in matrix n, and the step is to calculate the maximum row of influence and each column in matrix n
Coefficient correlation；
S25：Update the residual error r respectively arranged in matrix n residual error r, wherein residual error rjCalculation formula be rj=nj-λi*tj, should
Step is that all row in matrix n are left out into the maximum row of influence so that the result that each column is obtained is minimum (two norms)；
S26：By the residual error r of all row in matrix njWith first threshold ξ1It is compared, if there is rj＜ ξ1, perform step
S27, otherwise return to step S23；
S27：Matrix n is updated to n=r, and deletes nj, dictionary ΨpIt is updated to Ψp=[Ψp, λi]；
It is that residual error after being left out by calculating each column is made comparisons with first threshold in step S26 and step S27, judges whether
It is the larger row of influence, if it is there can be dictionary Ψ as a characteristic atomicpIn, otherwise select again.
S28：By two norms of the matrix n after renewal and Second Threshold ξ2It is compared, if | | n | | ＜ ξ2, perform step
S29, otherwise return to step S22, wherein ξ1≤ξ2；
S29：By dictionary ΨpIt is orthogonalized processing (Smith's orthogonalization process can be used), output characteristic atom；
Step S28 and step S29 are to judge whether to have selected all characteristic atomics, if selected complete, by what is obtained
Dictionary ΨpIt is orthogonal to be exported as characteristic atomic, otherwise continue to select.
S210：The characteristic atomic combining target matrix m of output is subjected to OMP reconstruction calculations and obtains corresponding coefficient, by phase
Often two norms of row calculating are sized the coefficient answered, and N rows are effective before judging according to coefficient magnitude and error sets requirement,
Then the corresponding corresponding preceding N of characteristic atomic for judging to extract is arranged as final characteristic atomic output.
In the target identification method of the present invention, all kinds of clarifications of objective are obtained by characteristic atomic extracting method first former
Son, will diagonally rearrange every classification target dictionary respectively per classification target each characteristic atomic, then by the dictionary of all kinds of targets
The comprehensive dictionary of arranged in parallel composition, by the way that characteristic atomic is carried out into such particular arrangement, coefficient is obtained with reference to sparse coefficient
Figure, recognizes directly in coefficient data and counts, you can point door distinguishes how many each target of every class target with sorting out, so as to
To carry out the identifying processing of multiple targets simultaneously；Traditional target identification method is compared in addition, is entered without extracting area-of-interest
Row matching primitives carry out Classification and Identification, and directly just can obtain inhomogeneity target by setting specific dictionary is respectively at difference
Coefficient region, processing procedure is simple, and amount of calculation is small, improves the speed of target identification in the picture.Wherein the present invention herein in connection with
The principles such as the limited isometry in compressive sensing theory, set the characteristics of for required identification target with background environment ga s safety degree
Specific characteristic atomic extracting method has been counted, and by the way that characteristic atomic is formed into dictionary by rational combination, has used this dictionary
The coefficient matrix for calculating and obtaining is combined with calculation matrix, more preferable target identification effect can be obtained, it is possible to while being weighed
The preferable reconstruct image of structure quality, identification is counted with parallel computation and can be independent of each other with obtaining reconstructed image, the saving time.
Above content is to combine specific preferred embodiment further description made for the present invention, it is impossible to assert
The specific implementation of the present invention is confined to these explanations.For those skilled in the art, do not taking off
On the premise of from present inventive concept, some equivalent substitutes or obvious modification can also be made, and performance or purposes are identical, all should
When being considered as belonging to protection scope of the present invention.
Claims (10)
1. a kind of target identification method based on compressed sensing, it is characterised in that comprise the following steps：
S1：Obtain at least two classification target master sample figures；
S2：According to the master sample figure of all kinds of targets, the spy of all kinds of targets is obtained using characteristic atomic extracting method
Levy atom；
S3：Each described characteristic atomic of target described in every class is diagonally rearranged into the dictionary Ψ of target described in every class respectivelyp,
And the dictionary arranged in parallel of all kinds of targets is constituted to comprehensive dictionary Ψ；
S4：Sampling is compressed to original image x to be identified using calculation matrix Φ, the sampled signal y compressed；
S5：With reference to comprehensive dictionary Ψ, calculation matrix Φ and sampled signal y, the original to be identified is obtained by reconstruction calculations
The sparse coefficient θ of beginning image；
S6：Sparse coefficient θ progress is handled and obtains coefficient figure, line number and size according to residing for connected domain in the coefficient figure
Classification and Identification and counting are carried out, to realize the identification to all kinds of targets in the original image.
2. target identification method according to claim 1, it is characterised in that step S1 is specifically included：According to image or regard
Frequency known target at least described in two classes existed in flowing, extracts the target image of all kinds of targets, will be per target described in class
Multiple target images the master sample figure for obtaining target described in every class is weighted respectively.
3. target identification method according to claim 2, it is characterised in that wherein extract the target of all kinds of targets
Image is specially：The target figure of all kinds of targets is extracted using morphological image method or manual identified dividing method
Picture.
4. target identification method according to claim 1, it is characterised in that step S4 also include using sampled signal y as
Data storage is in memory.
5. target identification method according to claim 1, it is characterised in that step S5 is specifically included：According to comprehensive word
Allusion quotation Ψ and calculation matrix Φ product obtain perceiving matrix A=Φ * Ψ, with reference to sampled signal y, pass through the reconstruct meter of OMP algorithms
Calculate formula y=A* θ and calculate the sparse coefficient θ for obtaining the original image.
6. target identification method according to claim 1, it is characterised in that also including step S7：By sparse coefficient θ with it is comprehensive
The dictionary Ψ of conjunction is multiplied, the reconstructed image x gathered1=Ψ * θ.
7. the target identification method according to any one of claim 1 to 6, it is characterised in that the characteristic atomic in step S2
Extracting method specifically extracts the clarification of objective atom using MOD algorithms or K-SVD algorithms.
8. the target identification method according to any one of claim 1 to 6, it is characterised in that the characteristic atomic in step S2
Extracting method is specifically included：
S21：The objective matrix m of the master sample figure of all kinds of targets is inputted, and is initialized：Matrix n is initialized as n
=m, circulation mark i=0, the dictionary Ψ of the target initiallypFor empty set；
S22：Assignment circulation mark i=i+1；
S23：The maximum kth row of two norms in matrix n all row are found, the column element λ of extraction is used asi, wherein
S is matrix n columns；
S24：The column element λ that all row are arranged with kth in calculating matrix niOptimal ratio t, wherein
J refers to the jth row in matrix n；
S25：Update the residual error r respectively arranged in matrix n residual error r, residual error rjCalculation formula be rj=nj-λi*tj；
S26：By the residual error r of all row in matrix njWith first threshold ξ1It is compared, if there is rj＜ ξ1, step S27 is performed, it is no
Then return to step S23；
S27：Matrix n is updated to n=r, and deletes nj, dictionary ΨpIt is updated to Ψp=[Ψp, λi]；
S28：By two norms of the matrix n after renewal and Second Threshold ξ2It is compared, if | | n | | ＜ ξ2, step S29 is performed, it is no
Then return to step S22；
S29：By dictionary ΨpIt is orthogonalized processing, output characteristic atom.
9. target identification method according to claim 8, it is characterised in that also include S210 in step S2：By output
Characteristic atomic combining target matrix m carries out OMP reconstruction calculations and obtains corresponding coefficient, and by corresponding coefficient, often row calculates two norms
It is sized, effectively, then the corresponding feature for judging to extract is former for N rows before judging according to coefficient magnitude and error sets requirement
The corresponding preceding N of son is arranged as final characteristic atomic output.
10. target identification method according to claim 8, it is characterised in that orthogonalization process is specifically used in step S29
Smith's orthogonalization process；Preferably, ξ1≤ξ2。
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710548175.XA CN107273908B (en) | 2017-07-06 | 2017-07-06 | A kind of compressed sensing based target identification method |
| PCT/CN2017/098652 WO2019006835A1 (en) | 2017-07-06 | 2017-08-23 | Target recognition method based on compressed sensing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710548175.XA CN107273908B (en) | 2017-07-06 | 2017-07-06 | A kind of compressed sensing based target identification method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107273908A true CN107273908A (en) | 2017-10-20 |
| CN107273908B CN107273908B (en) | 2019-09-06 |
Family
ID=60073082
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710548175.XA Active CN107273908B (en) | 2017-07-06 | 2017-07-06 | A kind of compressed sensing based target identification method |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN107273908B (en) |
| WO (1) | WO2019006835A1 (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108280818A (en) * | 2018-01-19 | 2018-07-13 | 清华大学深圳研究生院 | A kind of fast target imaging method and system based on compressed sensing |
| CN109033963A (en) * | 2018-06-22 | 2018-12-18 | 王连圭 | The trans-regional human motion posture target identification method of multiple-camera video |
| CN110472576A (en) * | 2019-08-15 | 2019-11-19 | 西安邮电大学 | A kind of method and device for realizing mobile human body Activity recognition |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111796253B (en) * | 2020-09-01 | 2022-12-02 | 西安电子科技大学 | Radar target constant false alarm detection method based on sparse signal processing |
| CN112508089B (en) * | 2020-12-03 | 2023-10-31 | 国网山西省电力公司晋城供电公司 | Self-adaptive compressed sensing method for partial discharge signal compressed transmission |
| CN113093164B (en) * | 2021-03-31 | 2023-02-10 | 西安电子科技大学 | Translation-invariant and noise-robust radar image target identification method |
| CN113670435B (en) * | 2021-08-20 | 2023-06-23 | 西安石油大学 | Underground vibration measuring device and method based on compressed sensing technology |
| CN113922823B (en) * | 2021-10-29 | 2023-04-21 | 电子科技大学 | Social media information propagation graph data compression method based on constraint sparse representation |
| CN116541470B (en) * | 2023-07-07 | 2024-02-13 | 深圳创维智慧科技有限公司 | Synchronization method, device, equipment and medium of database read-only standby library |
| CN117041359B (en) * | 2023-10-10 | 2023-12-22 | 北京安视华业科技有限责任公司 | Efficient compression transmission method for information data |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104574450A (en) * | 2014-12-31 | 2015-04-29 | 南京邮电大学 | Image reconstruction method based on compressed sensing |
| CN105631478A (en) * | 2015-12-25 | 2016-06-01 | 天津科技大学 | Plant classification method based on sparse expression dictionary learning |
| US20160157746A1 (en) * | 2014-12-01 | 2016-06-09 | The Regents Of The University Of California | Diffusion reproducibility evaluation and measurement (dream)-mri imaging methods |
| CN106203374A (en) * | 2016-07-18 | 2016-12-07 | 清华大学深圳研究生院 | A kind of characteristic recognition method based on compressed sensing and system thereof |
| CN106203453A (en) * | 2016-07-18 | 2016-12-07 | 清华大学深圳研究生院 | A kind of based on compressed sensing biological with abiotic target identification method and system thereof |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105844635B (en) * | 2016-03-21 | 2018-10-12 | 北京工业大学 | A kind of rarefaction representation depth image method for reconstructing based on structure dictionary |
| CN106157232B (en) * | 2016-06-30 | 2019-04-26 | 广东技术师范学院 | A kind of general steganalysis method of digital picture characteristic perception |
| CN106557784B (en) * | 2016-11-23 | 2020-05-01 | 上海航天控制技术研究所 | Rapid target identification method and system based on compressed sensing |
| CN106815806B (en) * | 2016-12-20 | 2020-01-10 | 浙江工业大学 | Single image SR reconstruction method based on compressed sensing and SVR |
-
2017
- 2017-07-06 CN CN201710548175.XA patent/CN107273908B/en active Active
- 2017-08-23 WO PCT/CN2017/098652 patent/WO2019006835A1/en active Application Filing
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20160157746A1 (en) * | 2014-12-01 | 2016-06-09 | The Regents Of The University Of California | Diffusion reproducibility evaluation and measurement (dream)-mri imaging methods |
| CN104574450A (en) * | 2014-12-31 | 2015-04-29 | 南京邮电大学 | Image reconstruction method based on compressed sensing |
| CN105631478A (en) * | 2015-12-25 | 2016-06-01 | 天津科技大学 | Plant classification method based on sparse expression dictionary learning |
| CN106203374A (en) * | 2016-07-18 | 2016-12-07 | 清华大学深圳研究生院 | A kind of characteristic recognition method based on compressed sensing and system thereof |
| CN106203453A (en) * | 2016-07-18 | 2016-12-07 | 清华大学深圳研究生院 | A kind of based on compressed sensing biological with abiotic target identification method and system thereof |
Non-Patent Citations (1)
| Title |
|---|
| 杨森林: "基于稀疏性估计的图像压缩感知研究", 《西安文理学院学报( 自然科学版)》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108280818A (en) * | 2018-01-19 | 2018-07-13 | 清华大学深圳研究生院 | A kind of fast target imaging method and system based on compressed sensing |
| CN108280818B (en) * | 2018-01-19 | 2020-04-03 | 清华大学深圳研究生院 | Rapid target imaging method and system based on compressed sensing |
| CN109033963A (en) * | 2018-06-22 | 2018-12-18 | 王连圭 | The trans-regional human motion posture target identification method of multiple-camera video |
| CN109033963B (en) * | 2018-06-22 | 2021-07-06 | 王连圭 | Multi-camera video cross-region human motion posture target recognition method |
| CN110472576A (en) * | 2019-08-15 | 2019-11-19 | 西安邮电大学 | A kind of method and device for realizing mobile human body Activity recognition |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2019006835A1 (en) | 2019-01-10 |
| CN107273908B (en) | 2019-09-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107273908B (en) | A kind of compressed sensing based target identification method | |
| CN108491849A (en) | Hyperspectral image classification method based on three-dimensional dense connection convolutional neural networks | |
| CN105005765B (en) | A kind of facial expression recognizing method based on Gabor wavelet and gray level co-occurrence matrixes | |
| CN108764072A (en) | A kind of blood cell subsets image classification method based on Multiscale Fusion | |
| CN103400154B (en) | A kind of based on the human motion recognition method having supervision Isometric Maps | |
| CN107424145A (en) | The dividing method of nuclear magnetic resonance image based on three-dimensional full convolutional neural networks | |
| CN101004791A (en) | Method for recognizing facial expression based on 2D partial least square method | |
| CN109389585A (en) | A kind of brain tissue extraction method based on full convolutional neural networks | |
| CN107622225B (en) | Face recognition method based on independent component analysis network | |
| CN109325507A (en) | A kind of image classification algorithms and system of combination super-pixel significant characteristics and HOG feature | |
| CN107092884A (en) | Rapid coarse-fine cascade pedestrian detection method | |
| CN110991554B (en) | Improved PCA (principal component analysis) -based deep network image classification method | |
| CN106803105B (en) | Image classification method based on sparse representation dictionary learning | |
| CN114842507A (en) | Reinforced pedestrian attribute identification method based on group optimization reward | |
| Liu et al. | Image retrieval using CNN and low-level feature fusion for crime scene investigation image database | |
| Asyhar et al. | Implementation LSTM Algorithm for Cervical Cancer using Colposcopy Data | |
| Monadjemi et al. | Experiments on high resolution images towards outdoor scene classification | |
| CN110516615A (en) | Human and vehicle shunting control method based on convolutional neural networks | |
| Gavhale et al. | Identification of medicinal plant using Machine learning approach | |
| Yusof et al. | Tropical wood species recognition system based on gabor filter as image multiplier | |
| Busch et al. | Texture classification using multiple wavelet analysis | |
| CN116630964A (en) | Food image segmentation method based on discrete wavelet attention network | |
| CN109165551B (en) | Expression recognition method for adaptively weighting and fusing significance structure tensor and LBP characteristics | |
| Zhao et al. | Pornographic image recognition in compressed domain based on multi-cost sensitive decision tree | |
| Divyasree et al. | Ayurvedic leaf identification using deep learning model: VGG16 |
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 |