CN110322011A - The object relationship building method and device of oriented inference model - Google Patents
The object relationship building method and device of oriented inference model Download PDFInfo
- Publication number
- CN110322011A CN110322011A CN201810267096.6A CN201810267096A CN110322011A CN 110322011 A CN110322011 A CN 110322011A CN 201810267096 A CN201810267096 A CN 201810267096A CN 110322011 A CN110322011 A CN 110322011A
- Authority
- CN
- China
- Prior art keywords
- combination
- data
- combined
- image
- natural language
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/082—Learning methods modifying the architecture, e.g. adding, deleting or silencing nodes or connections
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/04—Inference or reasoning models
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the invention provides a kind of object relationship building method of oriented inference model and devices, in this method, object extraction is carried out respectively based on different inference patterns for different data sources, obtain corresponding target object, then the corresponding target object of Various types of data is combined by default rule, by obtained combination to as input data to be trained, to solve the object relationship construction problem of multiple types mixing source data, a kind of make of simple and effective complex object relationship easily realized is provided for the training of inference pattern, be conducive to improve the joint training efficiency of different data sources datum target.
Description
Technical field
The present embodiments relate to field of computer technology, and in particular to a kind of object relationship construction of oriented inference model
Method and device.
Background technique
The knowledge that it is acquired can be applied to the task of general digital world by trained neural network --- knows
Other image, identification voice, the detection various applications such as disease or recommended advertisements.Neural network it is this faster more efficient
Mode can derive its new data obtained based on the content that it is trained, that is, reasoning.Reasoning is without training
Also it can generate, and reasoning task does not often need all infrastructure of its training program can reach good effect.
The target of training (be similar to the mankind and receive education) is knowledge acquisition, and the training of neural network and the mankind receive the process of education
There are relatively big differences.Neuron in human brain may be coupled to any other neuron in specific physical distance, and people
Artificial neural networks are divided into many different layers (layer), connection (connection) and data dissemination as being not
The direction of (data propagation).It is substantially a bulky galactic database by correctly trained neural network.
People's its all training process of having nothing for it but prepares all the things.If we want in real world using these training
As a result, we are desirable for a kind of quick application that can keep learning and capable of being applied to the data that it has never seen.
Here it is reasonings: only needing very small amount of real world data, can quickly obtain same correct answer.
However, during realizing innovation and creation, inventors have found that at present in artificial neural network mainly using
Deep learning model.And a deep learning model is " simple " continuous geometric transformation chain, and a vector space is reflected
It is mapped to another space.That it can do is exactly flag data X, and correspondence is associated on data Y.Only exist and learns from X to Y
Continuous transformation, and there is intensive available XY training set, deep learning model can be set up.That is, deep learning
The condition that model is set up is complex, and applicability is poor.And the training of inference pattern belong in artificial neural network one compared with
How new direction is based on inference pattern to training data at present, specifically how constructs to the data source of various structures,
It is still a urgent problem to be solved.
Summary of the invention
The embodiment of the present invention provides a kind of object relationship building method and device for oriented inference model.
In a first aspect, the embodiment of the present invention provides a kind of object relationship building method of oriented inference model,
For each data source in the mixing sample of input, target object is carried out based on corresponding inference pattern respectively
It extracts, obtains the depth representing of the corresponding target object of various data sources；
Based on preset rule of combination, the corresponding target object of variety classes data source is combined, and will be combined
To combination to as to be input to the input data being trained in multilayer neural network.
Second aspect, the embodiment of the present invention provide a kind of object relationship constructing apparatus of oriented inference model, comprising:
Multiple target object extracting unit, for being based on corresponding to each data source in the mixing sample for input
Inference pattern carry out target object extraction respectively, obtain the depth representing of the corresponding target object of various data sources；
Object relationship composite construction unit, for being based on preset rule of combination, by the corresponding mesh of variety classes data source
Mark object is combined, and the combination that combination is obtained is to as to be input to the input number being trained in multilayer neural network
According to.
The third aspect, another embodiment of the present invention provide a kind of computer equipment, including memory, processor and
On a memory and the computer program that can run on a processor, the processor realizes such as the when executing described program for storage
The step of one side the method.
Fourth aspect, another embodiment of the present invention provide a kind of computer readable storage medium, are stored thereon with meter
Calculation machine program, when which is executed by processor realize as described in relation to the first aspect method the step of.
The embodiment of the invention provides a kind of object relationship building method of oriented inference model and device, in this method,
Object extraction is carried out based on different inference patterns for different data sources respectively, obtains corresponding target object, then will
The corresponding target object of Various types of data is combined by default rule, by obtained combination to as input number to be trained
According to, thus solve multiple types mixing source data object relationship construction problem, provide a kind of letter for the training of inference pattern
The make of single complex object relationship effectively easily realized is conducive to the joint training effect for improving different data sources datum target
Rate.
Detailed description of the invention
By reading the following detailed description of the preferred embodiment, various other advantages and benefits are common for this field
Technical staff will become clear.The drawings are only for the purpose of illustrating a preferred embodiment, and is not considered as to the present invention
Limitation.And throughout the drawings, the same reference numbers will be used to refer to the same parts.In the accompanying drawings:
Fig. 1 is a kind of object relationship building method flow chart of oriented inference model provided in an embodiment of the present invention；
Fig. 2 is multiple target object extraction schematic diagram provided in an embodiment of the present invention；
Fig. 3 is object composition construction provided in an embodiment of the present invention and training schematic diagram；
Fig. 4 is a kind of object relationship constructing apparatus example structure schematic diagram of oriented inference model provided by the invention；
Fig. 5 is a kind of network side equipment example structure block diagram provided by the invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
In a first aspect, the embodiment of the invention provides a kind of object relationship building method of oriented inference model, such as Fig. 1 institute
Show, comprising:
S101, for each data source in the mixing sample of input, mesh is carried out based on corresponding inference pattern respectively
Object extraction is marked, the depth representing of the corresponding target object of various data sources is obtained；
S102, it is based on preset rule of combination, the corresponding target object of variety classes data source is combined, and by group
Obtained combination is closed to as to be input to the input data being trained in multilayer neural network.
In the object relationship building method of oriented inference model provided in an embodiment of the present invention, for different data source bases
Object extraction is carried out respectively in different inference patterns, obtains corresponding target object, then by the corresponding target of Various types of data
Object is combined by default rule, by obtained combination to as input data to be trained, to solve multiple types
Mix the object relationship construction problem of source data, for inference pattern training provide it is a kind of it is simple and effective easily realize it is complicated right
As the make of relationship, be conducive to the joint training efficiency for improving different data sources datum target.
In some embodiments, the form of the data source of the mixing sample in step S101 can be image, natural language
The combination of speech, voice etc. these data sources.Inference pattern in step S101 can be selected according to the type of data source, example
Such as when data source is image data, inference pattern can be convolutional neural networks (CNN) model based on deep learning；Work as number
According to source be natural language data when, inference pattern can for shot and long term remember (LSTM) model treatment.Here treatment process can
To specifically include: multi-target detection, target position detection, extraction of semantics etc. can export every kind after inference pattern is handled
The depth representing of type target object.
Fig. 2 shows the schematic diagrames that the target object of step S101 extracts, using the data source of mixing sample as image data
And natural language data instance.For the image data in the mixing sample of input, each target object (such as city is handed over
Street flow schematic diagram in logical, the image objects such as road, traffic lights, pedestrian, motor vehicle in automatic Pilot) it can be extracted
As a relatively independent unit obj (x).For the natural language data in the mixing sample of input, each and figure
As the related target object (such as urban weather, the text objects such as crowded type, great festivals or holidays activity notification) of data
It will be dealt with into character representation kw (x).After treatment to get several image object objects obj arrived as shown in the right side of fig 2
(x) and several natural language target object kw (x).
In some embodiments, corresponding for variety classes data source target object is combined in step S102
There are many kinds of mode is possible.It is understood that the type of the data source mixed in mixing sample is different, target object combination
Mode may also be different.A kind of optional mesh when mixing sample includes image data and natural language data is described below
The embodiment for marking object composition, specifically includes:
S2011, all corresponding image object objects of image data are subjected to combination of two, obtain several images two
Grade combination pair；
S2012, all corresponding natural language target objects of natural language data are subjected to combination of two, obtained several
A natural language second level combination pair；
S2013, it combines image second level combine pair to being combined to obtain to combine with natural language second level at random.
It specifically may refer to Fig. 3, obtain image object object obj (1), obj (2), obj (3), obj into step S101 is crossed
And natural language target object kw (1), kw (2), kw (3), kw (4) ... (4) ....By all image object objects
Obj (x) is combined at random two-by-two, obtains the combination pair of several image second levels；By all natural language target object kw (x)
It is combined at random two-by-two, obtains the combination pair of several natural language second levels.It then will be at random by an image second level combination pair
Combine with a natural language second level to being combined, formed one combination pair, the combination to it is inner include four target objects, should
Combination is to can be used as input data and be input in multilayer neural network to be trained.
Further, for including image data in mixing sample the case where, if the target object of certain combination centering is special
Sign is not significant enough, trains difficulty larger, in order to obtain preferable training effect, after step S2013, the embodiment of the present invention
The method of offer can also include:
S2014, specified raw image data is added to each combination centering；Wherein, specified raw image data
Are as follows: any one in source image data corresponding to all image object objects of the combination centering.
It specifically may refer to Fig. 3, that is to say, that if the image object object obj (x) that a combination centering includes is extracted from
Image data in a certain original mixed sample, then the raw image data can be added to the tectonic association centering to improve
The combination is to the conspicuousness for indicating feature.It, can if there is image object object obj (x) to extract from multiple raw image datas
It is added to combination centering to randomly choose a raw image data.
Mixing sample is extracted and relationship construction after, method provided in an embodiment of the present invention can also include
S103, the combination is trained to being input in multilayer neural network；
S104, the result that each layer neural metwork training exports is weighted expression, obtains the single of the mixing sample
It indicates.
Same to participate in Fig. 3, trained process is exactly to obtain the unification of target object combination using multilayer neural network in fact
Character representation.Specifically, firstly, obtaining a set, which may include the spy of combination pair obtained in all S102
Sign indicates.Then, all combined character representation matrixes of training output are added by element using the method for weighting.So far,
One group of mixing sample of input is somebody's turn to do by target object composite construction, multilayer neural network training, by element fusion output
The single representation of mixing sample.
Second aspect, the embodiment of the invention provides the object relationship constructing apparatus of another oriented inference model, such as Fig. 4
It is shown, comprising:
Multiple target object extracting unit 401, for each data source in the mixing sample for input, based on pair
The inference pattern answered carries out target object extraction respectively, obtains the depth representing of the corresponding target object of various data sources；
Object relationship composite construction unit 402, it is for being based on preset rule of combination, variety classes data source is corresponding
Target object is combined, and the combination that combination is obtained is to as to be input to the input being trained in multilayer neural network
Data.
It in some embodiments, include image data and natural language data in the mixing sample；
The object relationship composite construction unit 402 is based on preset rule of combination, and variety classes data source is corresponding
Target object is combined, comprising:
All corresponding image object objects of image data are subjected to combination of two, obtain the combination of several image second levels
It is right；
All corresponding natural language target objects of natural language data are subjected to combination of two, obtain several natures
The combination pair of language second level；
It at random combines image second level and combines pair to being combined to obtain to combine with natural language second level.
In some embodiments, the object relationship composite construction unit 402, at random by image second level combine to
The combination of natural language second level is combined to later to being combined to obtain, and is also used to:
Specified raw image data is added to each combination centering；Wherein, the specified original image number
According to are as follows: any one in source image data corresponding to all image object objects of the combination centering.
In some embodiments, described device further include:
Object relationship combined training unit 403, for will the obtained combination of combination to be input in multilayer neural network into
Row training；
Training integrated unit 404, the result for exporting each layer neural metwork training are weighted expression, obtain described
The single representation of mixing sample.
The object relationship constructing apparatus for the oriented inference model introduced by second aspect be can execute it is of the invention real
The device of the object relationship building method of the oriented inference model in example is applied, so based on face described in the embodiment of the present invention
The method constructed to the object relationship of inference pattern, those skilled in the art can understand the oriented inference mould of the present embodiment
The specific embodiment of the object relationship constructing apparatus of type and its various change form, so herein for the oriented inference mould
How the object relationship constructing apparatus of type realizes the object relationship building method of the oriented inference model in the embodiment of the present invention not
It is discussed in detail again.As long as those skilled in the art implement the object relationship construction of oriented inference model in the embodiment of the present invention
Device used by method belongs to the range to be protected of the application.
Fig. 5 shows the structural block diagram of network side equipment provided in an embodiment of the present invention.
Referring to Fig. 5, the network side equipment, comprising: processor (processor) 501, memory (memory) 502 and
Bus 503；
Wherein, the processor 501 and memory 502 complete mutual communication by the bus 503.
The processor 501 is used to call the program instruction in the memory 502, to execute first aspect embodiment institute
The method of offer.
A kind of computer program product is also disclosed in the embodiment of the present invention, and the computer program product is non-temporary including being stored in
Computer program on state computer readable storage medium, the computer program include program instruction, when described program instructs
When being computer-executed, computer is able to carry out method provided by above-mentioned first aspect embodiment.
The embodiment of the present invention also provides a kind of non-transient computer readable storage medium, and the non-transient computer is readable to deposit
Storage media stores computer instruction, and the computer instruction executes the computer provided by above-mentioned first aspect embodiment
Method.
In the instructions provided here, numerous specific details are set forth.It is to be appreciated, however, that implementation of the invention
Example can be practiced without these specific details.In some instances, well known method, structure is not been shown in detail
And technology, so as not to obscure the understanding of this specification.
Similarly, it should be understood that in order to simplify the disclosure and help to understand one or more of the various inventive aspects,
Above in the description of exemplary embodiment of the present invention, each feature of the invention is grouped together into single implementation sometimes
In example, figure or descriptions thereof.However, the disclosed method should not be interpreted as reflecting the following intention: i.e. required to protect
Shield the present invention claims features more more than feature expressly recited in each claim.More precisely, as following
Claims reflect as, inventive aspect is all features less than single embodiment disclosed above.Therefore,
Thus the claims for following specific embodiment are expressly incorporated in the specific embodiment, wherein each claim itself
All as a separate embodiment of the present invention.
Those skilled in the art will understand that can be carried out adaptively to the module in the equipment in embodiment
Change and they are arranged in one or more devices different from this embodiment.It can be the module or list in embodiment
Member or component are combined into a module or unit or component, and furthermore they can be divided into multiple submodule or subelement or
Sub-component.Other than such feature and/or at least some of process or unit exclude each other, it can use any
Combination is to all features disclosed in this specification (including adjoint claim, abstract and attached drawing) and so disclosed
All process or units of what method or apparatus are combined.Unless expressly stated otherwise, this specification is (including adjoint power
Benefit require, abstract and attached drawing) disclosed in each feature can carry out generation with an alternative feature that provides the same, equivalent, or similar purpose
It replaces.
In addition, it will be appreciated by those of skill in the art that although some embodiments in this include institute in other embodiments
Including certain features rather than other feature, but the combination of the feature of different embodiment means in the scope of the present invention
Within and form different embodiments.For example, in the following claims, embodiment claimed it is any it
One can in any combination mode come using.
Certain unit embodiments of the invention can be implemented in hardware, or to run on one or more processors
Software module realize, or be implemented in a combination thereof.It will be understood by those of skill in the art that can be used in practice
Microprocessor or digital signal processor (DSP) realize gateway according to an embodiment of the present invention, proxy server, in system
Some or all components some or all functions.The present invention is also implemented as executing side as described herein
Some or all device or device programs (for example, computer program and computer program product) of method.It is such
It realizes that program of the invention can store on a computer-readable medium, or can have the shape of one or more signal
Formula.Such signal can be downloaded from an internet website to obtain, and perhaps be provided on the carrier signal or with any other shape
Formula provides.
It should be noted that the above-mentioned embodiments illustrate rather than limit the invention, and ability
Field technique personnel can be designed alternative embodiment without departing from the scope of the appended claims.In the claims,
Any reference symbol between parentheses should not be configured to limitations on claims.Word "comprising" does not exclude the presence of not
Element or step listed in the claims.Word "a" or "an" located in front of the element does not exclude the presence of multiple such
Element.The present invention can be by means of including the hardware of several different elements and being come by means of properly programmed computer real
It is existing.In the unit claims listing several devices, several in these devices can be through the same hardware branch
To embody.The use of word first, second, and third does not indicate any sequence.These words can be explained and be run after fame
Claim.
Claims (10)
1. a kind of object relationship building method of oriented inference model characterized by comprising
For each data source in the mixing sample of input, carries out target object respectively based on corresponding inference pattern and mention
It takes, obtains the depth representing of the corresponding target object of various data sources；
Based on preset rule of combination, the corresponding target object of variety classes data source is combined, and combination is obtained
Combination is to as to be input to the input data being trained in multilayer neural network.
2. the method according to claim 1, wherein including image data and natural language in the mixing sample
Say data；
It is described to be based on preset rule of combination, the corresponding target object of variety classes data source is combined, comprising:
All corresponding image object objects of image data are subjected to combination of two, obtain the combination pair of several image second levels；
All corresponding natural language target objects of natural language data are subjected to combination of two, obtain several natural languages
Second level combination pair；
It at random combines image second level and combines pair to being combined to obtain to combine with natural language second level.
3. according to the method described in claim 2, it is characterized in that, at random by image second level combine to natural language second level
After combination is to the step for being combined to obtain combination pair, the method also includes:
Specified raw image data is added to each combination centering；Wherein, the specified raw image data are as follows:
Any one in source image data corresponding to all image object objects of the combination centering.
4. the method according to claim 1, wherein the method also includes:
The combination is trained to being input in multilayer neural network；
The result that each layer neural metwork training exports is weighted expression, obtains the single representation of the mixing sample.
5. a kind of object relationship constructing apparatus of oriented inference model characterized by comprising
Multiple target object extracting unit, for being pushed away based on corresponding to each data source in the mixing sample for input
Reason model carries out target object extraction respectively, obtains the depth representing of the corresponding target object of various data sources；
Object relationship composite construction unit, for being based on preset rule of combination, by the corresponding target pair of variety classes data source
As being combined, and the combination that combination is obtained is to as to be input to the input data being trained in multilayer neural network.
6. device according to claim 5, which is characterized in that include image data and natural language in the mixing sample
Say data；
The object relationship composite construction unit is based on preset rule of combination, by the corresponding target object of variety classes data source
It is combined, comprising:
All corresponding image object objects of image data are subjected to combination of two, obtain the combination pair of several image second levels；
All corresponding natural language target objects of natural language data are subjected to combination of two, obtain several natural languages
Second level combination pair；
It at random combines image second level and combines pair to being combined to obtain to combine with natural language second level.
7. device according to claim 6, which is characterized in that the object relationship composite construction unit will scheme at random
As second level combination is combined to later to being combined to obtain to combining with natural language second level, it is also used to:
Specified raw image data is added to each combination centering；Wherein, the specified raw image data are as follows:
Any one in source image data corresponding to all image object objects of the combination centering.
8. device according to claim 5, which is characterized in that described device further include:
Object relationship combined training unit, for the obtained combination of combination to be trained to being input in multilayer neural network；
Training integrated unit, the result for exporting each layer neural metwork training are weighted expression, obtain the aggregate sample
This single representation.
9. a kind of computer equipment, can run on a memory and on a processor including memory, processor and storage
Computer program, which is characterized in that the processor is realized when executing described program such as any the method for claim 1-4
Step.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the program is by processor
The step of the method as any such as claim 1-4 is realized when execution.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810267096.6A CN110322011B (en) | 2018-03-28 | 2018-03-28 | Inference model-oriented object relationship construction method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810267096.6A CN110322011B (en) | 2018-03-28 | 2018-03-28 | Inference model-oriented object relationship construction method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110322011A true CN110322011A (en) | 2019-10-11 |
| CN110322011B CN110322011B (en) | 2021-08-06 |
Family
ID=68110412
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810267096.6A Expired - Fee Related CN110322011B (en) | 2018-03-28 | 2018-03-28 | Inference model-oriented object relationship construction method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110322011B (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110087481A1 (en) * | 2009-10-14 | 2011-04-14 | Institute For Information Industry | Translation system, translation method and computer readable-writable storage medium of the same |
| CN102999640A (en) * | 2013-01-09 | 2013-03-27 | 公安部第三研究所 | Video and image retrieval system and method based on semantic reasoning and structural description |
| CN104036257A (en) * | 2014-06-25 | 2014-09-10 | 华南农业大学 | Citrus vein phloem degeneration detection and classification method based on D-S theory through multi-source data fusion |
| CN104142921A (en) * | 2013-05-06 | 2014-11-12 | 苏州普达新信息技术有限公司 | Image feature optimal-combination method based on cloud computing |
| CN105184308A (en) * | 2015-08-03 | 2015-12-23 | 北京航空航天大学 | Remote sensing image building detection and classification method based on global optimization decision |
| CN105426356A (en) * | 2015-10-29 | 2016-03-23 | 杭州九言科技股份有限公司 | Target information identification method and apparatus |
| CN106021364A (en) * | 2016-05-10 | 2016-10-12 | 百度在线网络技术（北京）有限公司 | Method and device for establishing picture search correlation prediction model, and picture search method and device |
| CN106547880A (en) * | 2016-10-26 | 2017-03-29 | 重庆邮电大学 | A kind of various dimensions geographic scenes recognition methodss of fusion geographic area knowledge |
| US20170330079A1 (en) * | 2016-05-11 | 2017-11-16 | International Business Machines Corporation | Automated Distractor Generation by Performing Disambiguation Operations |
| CN107480206A (en) * | 2017-07-25 | 2017-12-15 | 杭州电子科技大学 | A kind of picture material answering method based on multi-modal low-rank bilinearity pond |
-
2018
- 2018-03-28 CN CN201810267096.6A patent/CN110322011B/en not_active Expired - Fee Related
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110087481A1 (en) * | 2009-10-14 | 2011-04-14 | Institute For Information Industry | Translation system, translation method and computer readable-writable storage medium of the same |
| CN102999640A (en) * | 2013-01-09 | 2013-03-27 | 公安部第三研究所 | Video and image retrieval system and method based on semantic reasoning and structural description |
| CN104142921A (en) * | 2013-05-06 | 2014-11-12 | 苏州普达新信息技术有限公司 | Image feature optimal-combination method based on cloud computing |
| CN104036257A (en) * | 2014-06-25 | 2014-09-10 | 华南农业大学 | Citrus vein phloem degeneration detection and classification method based on D-S theory through multi-source data fusion |
| CN105184308A (en) * | 2015-08-03 | 2015-12-23 | 北京航空航天大学 | Remote sensing image building detection and classification method based on global optimization decision |
| CN105426356A (en) * | 2015-10-29 | 2016-03-23 | 杭州九言科技股份有限公司 | Target information identification method and apparatus |
| CN106021364A (en) * | 2016-05-10 | 2016-10-12 | 百度在线网络技术（北京）有限公司 | Method and device for establishing picture search correlation prediction model, and picture search method and device |
| US20170330079A1 (en) * | 2016-05-11 | 2017-11-16 | International Business Machines Corporation | Automated Distractor Generation by Performing Disambiguation Operations |
| CN106547880A (en) * | 2016-10-26 | 2017-03-29 | 重庆邮电大学 | A kind of various dimensions geographic scenes recognition methodss of fusion geographic area knowledge |
| CN107480206A (en) * | 2017-07-25 | 2017-12-15 | 杭州电子科技大学 | A kind of picture material answering method based on multi-modal low-rank bilinearity pond |
Non-Patent Citations (2)
| Title |
|---|
| ZHONGLI MA等: "Multi-targets recognition for surface moving platform vision system based on combined features", 《2014 IEEE INTERNATIONAL CONFERENCE ON MECHATRONICS AND AUTOMATION》 * |
| 张超鹏: "基于稀疏表示和混合样本的图像超分辨率重建算法研究", 《中国优秀硕士学位论文全文数据库》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110322011B (en) | 2021-08-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Brown et al. | Engaging Local Communities in Stewardship of World Heritage: A methodology based on the COMPACT experience | |
| CN110147456A (en) | A kind of image classification method, device, readable storage medium storing program for executing and terminal device | |
| Evers et al. | Knowledge cluster formation in Peninsular Malaysia: The emergence of an epistemic landscape | |
| Syakur et al. | Ensuring local stakeholder support for marine conservation: establishing a locally-managed marine area network in Aceh | |
| Cai et al. | How does ICT expansion drive “smart” urban growth? A case study of Nanjing, China | |
| CN110322011A (en) | The object relationship building method and device of oriented inference model | |
| Eglash et al. | Fractals, complexity, and connectivity in Africa | |
| Trautmann et al. | A classic problem | |
| Touya | Relevant space partitioning for collaborative generalisation | |
| Zou et al. | The Transformation of Coastal Governance, from Human Ecology to Local State, in the Jimei Peninsula, Xiamen, China | |
| Ramírez‐Monsalve et al. | Marine restoration governance arrangements: Issues of legitimacy | |
| Steingart | Law, computing and redistricting in the 1960s | |
| Kourri | A Tunnel of Many Worlds | |
| De Zeeuw | Improving Non-Observational Experiences: Channelling and Ordering. | |
| Santa et al. | 8 Protected Area Tourism in Batanes Islands, Philippines: Issues and Impacts | |
| Rowe | Towards an alternative theoretical framework for understanding local economic development | |
| CN110322037A (en) | Method for predicting and device based on inference pattern | |
| Lüscher | Characterising urban space from topographic databases: cartographic pattern recognition based on semantic modelling | |
| Van Cong et al. | Integrated coastal zone management in Vietnam status and emerging problems | |
| Gabriel | Pyow-Hack: Ordered Compositions in Lewis-Skyrms Signaling Games | |
| Jacuniak-Suda et al. | Governance arrangements in the Hamburg Metropolitan Region: between hard and soft institutional spaces | |
| Henocque et al. | Eu’s strategy on maritime & environmental issues in the four seas: multilateral approaches in the baltic, black, caspian & mediterranean seas | |
| Al Khalifa | An urban healing agenda for reform in Bahrain: where the dweller falls into the urban gap and the sailing boat hits the skyscraper | |
| Schothorst | The Power of Maps | |
| Vries | Integrating knowledge and policy in an uncertain world: the case of Offshore Wind Farm planning in the Dutch part of the North Sea. |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20210806 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |