CN109471959B - Figure reasoning model-based method and system for identifying social relationship of people in image - Google Patents
Figure reasoning model-based method and system for identifying social relationship of people in image Download PDFInfo
- Publication number
- CN109471959B CN109471959B CN201811646108.2A CN201811646108A CN109471959B CN 109471959 B CN109471959 B CN 109471959B CN 201811646108 A CN201811646108 A CN 201811646108A CN 109471959 B CN109471959 B CN 109471959B
- Authority
- CN
- China
- Prior art keywords
- graph
- social relationship
- network
- nodes
- semantic
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Devices For Executing Special Programs (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a figure social relationship identification method in an image based on a graph reasoning model, which comprises the following steps: step S1, modeling the coexistence relationship of the social relationship and the semantic object to form a priori knowledge graph; step S2, combining the high-level knowledge graph and the deep learning to form a graph inference network which can learn end to end, and using the propagation mechanism of the graph inference network to propagate and learn the information of the nodes in the prior knowledge graph to obtain weighted graph node output; and step S3, combining the weighted output of the graph nodes and outputting the social relationship label distribution of the target person pair through the classification network.
Description
Technical Field
The invention relates to the technical field of computer vision, in particular to a figure social relationship identification method and system in an image based on a graph reasoning model.
Background
Social relationships are fundamental components of social networks and play a very important role in daily life. In the era of people-oriented and artificial intelligence rapid development, intelligent analysis of social relationships has important applications in many fields, such as multi-target tracking, character track prediction, group activity analysis, human-computer interaction and the like.
The social relationship identification task has certain difficulty because the social relationship of people may completely change in different scenes. For example, if the target person pair is in the studio, they are more likely to be in a co-workers relationship, and if the target person pair is in the kitchen, they are more likely to be in a family relationship. This means that the target person plays an important role in the environment to identify the social relationship. However, the current method for solving the social relationship recognition task is only focused on the shape of the target person pair, or only extracts the candidate region of the target person pair, which may have related information around, to assist in the recognition of the social relationship, but ignores the semantic information of the object in the scene where the target person is located, and cannot accurately recognize the social relationship between the persons in the image.
Disclosure of Invention
In order to overcome the defects of the prior art, the present invention provides a method and a system for identifying a social relationship of a person in an image based on a graph-inference model, so as to automatically pay attention to an object graph node beneficial to identifying a current social relationship through a graph-inference mechanism, and ignore an object graph node not beneficial to identifying the social relationship, thereby further improving the efficiency of identifying the social relationship of the person in the image.
In order to achieve the above and other objects, the present invention provides a method for identifying a social relationship between people in an image based on a graph-based reasoning model, comprising the steps of:
step S1, modeling the coexistence relationship of the social relationship and the semantic object, and forming a priori knowledge graph;
step S2, combining the high-level knowledge graph and the deep learning to form a graph inference network which can learn end to end, and using the propagation mechanism of the graph inference network to propagate and learn the information of the nodes in the prior knowledge graph to obtain weighted graph node output;
and step S3, combining the weighted output of the graph nodes and outputting the social relationship label distribution of the target character pair through the classification network.
Preferably, in step S1, the social relationship training set is used to count the number of instances in which each social relationship and each semantic object coexist, so as to obtain a connection matrix of a graph structure with the social relationship and the semantic object as nodes.
Preferably, in step S1, according to the social relationship training set, the number of instances in which each social relationship and each semantic object coexist is counted, and the proportion of the instances is used as the value of the edge between the corresponding social relationship node and the semantic node in the graph structure, so as to form the connection matrix of the graph structure.
Preferably, in step S2, the graph inference network is used to extract the features of the target person pair and its surrounding semantic objects from the image containing the target person pair and its surrounding semantic objects, and the features are used as the input of the nodes corresponding to the graph inference network structure, and after passing through the learnable graph inference network, the weighted graph nodes are output.
Preferably, the graph inference network includes a convolutional neural network and a gate graph neural network, the graph inference network firstly uses the convolutional neural network to extract the target person pairs in the image containing the target person pairs and the semantic objects around the target person pairs and the features of the semantic objects obtained by the detector, then uses the propagation mechanism of the gate graph neural network to propagate and learn the mutual information among the graph nodes, and finally obtains the output O of all the graph nodes.
Preferably, the features of the target person pairs are used as initial values of social relationship nodes in the graph inference network, and the features of the semantic objects are used as initial values of corresponding semantic object nodes in the graph inference network.
Preferably, in step S3, the input of the classification network is the output O of the graph inference networkiThe output is social relationship label distribution.
Preferably, the classification network is a fully connected network, for eachInput OiObtaining a classification score si。
In order to achieve the above object, the present invention further provides a system for identifying a social relationship between people in an image based on a graph-based reasoning model, comprising:
the prior knowledge graph establishing unit is used for modeling the coexistence relationship of the social relationship and the semantic object to form a prior knowledge graph;
the graph inference unit is used for combining a high-level priori knowledge graph with deep learning to form a graph inference network capable of learning end to end, and the propagation mechanism in the graph inference network is used for propagating and learning the information of the nodes in the priori knowledge graph to obtain weighted graph node output;
and the classification unit is used for outputting the weighted graph nodes in a combined manner and outputting the social relationship label distribution of the target character pair through a classification network.
Preferably, the graph inference unit extracts the features of the target person pair and the semantic objects around the target person pair from the image including the target person pair and the semantic objects around the target person pair using the graph inference network, and obtains the output of the weighted graph nodes by using the features as the input of the nodes corresponding to the graph structure through the learnable graph inference network.
Compared with the prior art, the figure social relationship identification method and system based on the graph inference model, disclosed by the invention, form the prior knowledge graph by carrying out coexistence modeling on social relationships and semantic objects, fully fuse the information of the social relationships and the semantic objects through the knowledge graph structure, improve the characteristics to a new height, automatically pay attention to the object graph nodes beneficial to identifying the current social relationships through a graph inference mechanism, ignore the object graph nodes which are not beneficial to identifying the social relationships, and further improve the efficiency of identifying the person social relationships in the image.
Drawings
FIG. 1 is a flow chart of steps of a method for identifying a social relationship of a person in an image based on a graph inference model according to the invention;
FIG. 2 is a schematic diagram illustrating an architecture of an inference network in accordance with an embodiment of the present invention;
fig. 3 is a system architecture diagram of a system for identifying a social relationship of a person in an image based on a graph-inference model according to the present invention.
Detailed Description
Other advantages and capabilities of the present invention will be readily apparent to those skilled in the art from the present disclosure by describing the embodiments of the present invention with specific embodiments thereof in conjunction with the accompanying drawings. The invention is capable of other and different embodiments and its several details are capable of modification in various other respects, all without departing from the spirit and scope of the present invention.
FIG. 1 is a flowchart of the steps of a method for identifying a social relationship between people in an image based on a graph-based reasoning model according to the present invention. As shown in FIG. 1, the invention relates to a method for identifying a social relationship of a person in an image based on a graph reasoning model, which comprises the following steps:
and step S1, modeling the coexistence relationship of the social relationship and the semantic object to form a priori knowledge graph.
Specifically, in step S1, the social relationship training set is used to count the number of instances in which each social relationship and each semantic object coexist, so as to obtain the connection matrix M of the graph structure with the social relationship and the semantic object as nodes. In the embodiment of the present invention, the social relationship training set refers to a data set that includes many images and has label information, and may be an existing public data set, or a specific data set collected by itself, that is, in step S1, according to the social relationship training set, the number of instances where each social relationship and each semantic object coexist is counted, and the ratio of the number of instances occupied by the instances is used as the value of the edge between the corresponding social relationship node and the semantic node in the graph structure, so as to form the connection matrix M of the graph structure, where the social relationship is, for example, relatives, friends, and colleagues, and the semantic object refers to other general objects in the images except for the target person, such as a table, a chair, a water cup, and the like. For example, if a picture shows that two people are related to a colleague, the environment is a studio, and the studio has a computer, a keyboard, a desk, a tie, etc., then the objects count +1 accordingly for the example of the relationship of "colleague", count the whole data set, and convert the count into a frequency as the value of the corresponding position in the connection matrix.
And step S2, combining the prior knowledge graph of the high level with deep learning to form a graph inference network capable of learning end to end, propagating and learning the information of the nodes in the prior knowledge graph by using the propagation mechanism of the graph inference network to explore the interaction between the social relationship and the semantic objects and fully fuse the information of the social relationship and the semantic objects, specifically, extracting the characteristics of the target character pair and the semantic objects around the target character pair from the image containing the target character pair and the semantic objects around the target character pair by using the graph inference network, taking the characteristics as the input of the corresponding nodes of the graph structure, and obtaining the output of the weighted graph nodes after passing through the learnable graph inference network.
In the embodiment of the present invention, as shown in fig. 2, the graph inference network includes a Convolutional Neural Network (CNN) and a Gate Graph Neural Network (GGNN), first, the convolutional neural network is used to extract a target person pair in an image containing the target person pair and its surrounding semantic objects and the features of the semantic objects obtained by a detector, wherein the features of the target person pair are used as initial values of social relationship nodes in the graph inference network, and the features of the semantic objects are used as initial values of corresponding semantic object nodes in the graph inference network; then, the propagation mechanism of a Gate Graph Neural Network (GGNN) is utilized to propagate and learn mutual information among graph nodes, finally, the output O of all graph nodes is obtained, the importance of different semantic object nodes is inferred aiming at each social relationship by utilizing the gate graph neural network, and specifically, a relationship node set is { r }1，r2，...，rMThe semantic object node set is { o }1，o2，...，oNM and N respectively represent the number of corresponding node sets, correspond to each social relationship node i, and infer the importance alpha of semantic node jijWhere j ∈ Ni,NiRepresenting the adjacent node set of the node i to finally obtain the graph node output O corresponding to each relationshipi＝{oi，αi1oo1，αi1oo1，…αiNooNIn which o isi∈O,ooj∈O(j∈[1，N])。
And step S3, combining the weighted output of the graph nodes, and outputting the social relationship label distribution of the target character pair through the classification network. In an embodiment of the present invention, the input of the classification network is O in step S2iThe output is a social relationship label distribution, the classification network is a fully connected network, and for each input OiObtaining a classification score si。
Fig. 3 is a system architecture diagram of a system for identifying a social relationship of a person in an image based on a graph-inference model according to the present invention. As shown in fig. 3, the present invention provides a system for identifying a social relationship between people in an image based on a graph-based inference model, comprising:
the priori knowledge graph establishing unit 301 is configured to model a coexistence relationship between a social relationship and a semantic object, and form a priori knowledge graph.
Specifically, the priori knowledge graph creating unit 301 calculates the number of instances where each social relationship and each semantic object coexist by using the social relationship training set, thereby obtaining the connection matrix M of the graph structure using the social relationship and the semantic object as nodes. In the embodiment of the present invention, the priori knowledge graph establishing unit 301 calculates the number of instances where each social relationship and each semantic object coexist according to the social relationship training set, and uses the ratio of the number of instances as the value of the edge between the corresponding social relationship node and the semantic node in the graph structure, thereby forming the connection matrix M of the graph structure.
And the graph inference unit 302 is used for combining the high-level knowledge graph with deep learning to form a graph inference network capable of learning end to end, and propagating and learning the information of the nodes in the graph by using a propagation mechanism in the graph so as to explore the interaction between social relations and semantic objects. Specifically, the graph inference unit 302 extracts the features of the target person pair and the semantic objects around the target person pair from the image including the target person pair and the semantic objects around the target person pair using the graph inference network, and obtains the output of the weighted graph nodes by using the features as the input of the nodes corresponding to the graph structure through the learnable graph inference network.
In the specific embodiment of the present inventionFirstly, extracting a target person pair in an image containing the target person pair and semantic objects around the target person pair and the characteristics of the semantic objects obtained by a detector by using the convolutional neural network respectively, wherein the characteristics of the target person pair are used as initial values of social relation nodes in the graph inference network, and the characteristics of the semantic objects are used as initial values of corresponding semantic object nodes in the graph inference network; then, the propagation mechanism of a Gate Graph Neural Network (GGNN) is utilized to propagate and learn mutual information among graph nodes, finally, the output O of all graph nodes is obtained, the significance of different semantic object nodes is deduced by utilizing the gate graph neural network aiming at each social relation, and in particular, in the graph 3, t is the iteration number of the GGNN; a refers to the importance (weight) of each semantic object node obtained by inference of the graph inference network; f refers to the union characteristic of the graph nodes. Let the set of relational nodes be r1，r2，...，rM}, the semantic object node set is { o1，o2，...，oNM and N respectively represent the number of corresponding node sets, correspond to each social relationship node i, and infer the importance alpha of semantic node jijWhere j ∈ Ni,NiRepresenting the adjacent node set of the node i to finally obtain the graph node output O corresponding to each relationi＝{oi，αi1，oo1，αi1oo1，...αiNooNH, o thereini∈O,ooj∈O(j∈[1，N])
And the classification unit 303 is configured to output the social relationship label distribution of the target person pair through a classification network in combination with the weighted output of the graph nodes. In an embodiment of the present invention, the input of the classification network is the output O of the graph inference unit 302iThe output is a social relationship label distribution, the classification network is a fully connected network, and for each input OiObtaining a classification score si。
In summary, the method and system for identifying the social relationship of the person in the image based on the graph inference model form the prior knowledge graph by modeling the coexistence of the social relationship and the semantic objects, fully integrate the information of the social relationship and the semantic objects through the knowledge graph structure, improve the characteristics to a new height, automatically pay attention to the object graph nodes beneficial to identifying the current social relationship through the graph inference mechanism, ignore the object graph nodes which are not beneficial to identifying the social relationship, and further improve the efficiency of identifying the social relationship of the person in the image.
The foregoing embodiments are merely illustrative of the principles and utilities of the present invention and are not intended to limit the invention. Modifications and variations can be made to the above-described embodiments by those skilled in the art without departing from the spirit and scope of the present invention. Therefore, the protection scope of the present invention should be as set forth in the claims.
Claims (7)
1. A figure social relationship identification method in an image based on a graph reasoning model comprises the following steps:
step S1, modeling the coexistence relationship of the social relationship and the semantic object to form a priori knowledge graph;
step S2, combining the prior knowledge graph of the high layer with deep learning to form a graph inference network which can learn end to end, and using the propagation mechanism of the graph inference network to propagate and learn the information of the nodes in the prior knowledge graph to obtain weighted graph node output;
step S3, combining the weighted graph nodes and outputting the social relationship label distribution of the target character pair through the classification network;
in step S2, the graph inference network is used to extract the features of the target character pair and its surrounding semantic objects from the image containing the target character pair and its surrounding semantic objects, and the features are used as the input of the corresponding nodes of the graph inference network structure, and after passing through the learnable graph inference network, the weighted graph nodes are output;
the graph inference network comprises a convolutional neural network and a gate graph neural network, wherein the graph inference network firstly extracts a target person pair in an image containing the target person pair and semantic objects around the target person pair by using the convolutional neural network and the characteristics of the semantic objects acquired by a detector, then propagates and learns mutual information among graph nodes by using a propagation mechanism of the gate graph neural network, and finally obtains the output O of all the graph nodes.
2. The method for identifying the social relationship among the people in the image based on the graph-based reasoning model as claimed in claim 1, wherein: in step S1, the social relationship training set is used to count the number of instances where each social relationship and each semantic object coexist, so as to obtain a connection matrix of a graph structure with the social relationship and the semantic object as nodes.
3. The method for identifying the social relationship among the people in the image based on the graph-based reasoning model as claimed in claim 2, wherein: in step S1, the number of instances where each social relationship and each semantic object coexist is counted according to the social relationship training set, and the ratio of the number of instances in all the instances is used as the value of the edge between the corresponding social relationship node and the semantic node in the graph structure, thereby forming the connection matrix of the graph structure.
4. The method for identifying the social relationship among the people in the image based on the graph-based reasoning model as claimed in claim 1, wherein: the characteristics of the target character pairs are used as initial values of social relation nodes in the graph inference network, and the characteristics of the semantic objects are used as initial values of corresponding semantic object nodes in the graph inference network.
5. The method for identifying the social relationship among the people in the image based on the graph-based reasoning model as claimed in claim 1, wherein: in step S3, the input of the classification network is the output O of the graph inference networkiThe output is social relationship label distribution.
6. The method for identifying the social relationship among the people in the image based on the graph-based reasoning model as claimed in claim 5, wherein: the classification network is a fully connected network, for each input OiObtaining a classification score si。
7. A figure social relationship identification system in an image based on a graph reasoning model comprises:
the prior knowledge graph establishing unit is used for modeling the coexistence relationship of the social relationship and the semantic object to form a prior knowledge graph;
the graph inference unit is used for combining a high-level priori knowledge graph with deep learning to form an end-to-end learnable graph inference network, and the propagation mechanism in the graph inference network is used for propagating and learning the information of the nodes in the priori knowledge graph to obtain weighted graph node output;
the classification unit is used for outputting the weighted graph nodes in a combined manner, and outputting the social relation label distribution of the target character pair through a classification network;
the graph reasoning unit extracts the characteristics of the target character pair and the surrounding semantic objects of the target character pair from the image containing the target character pair and the surrounding semantic objects by utilizing the graph reasoning network, takes the characteristics as the input of the corresponding nodes of the graph structure, and obtains the output of weighted graph nodes after passing through the learnable graph reasoning network;
the graph inference network comprises a convolutional neural network and a gate graph neural network, wherein the graph inference network firstly extracts a target person pair in an image containing the target person pair and semantic objects around the target person pair by using the convolutional neural network and the characteristics of the semantic objects acquired by a detector, then propagates and learns mutual information among graph nodes by using a propagation mechanism of the gate graph neural network, and finally obtains the output O of all the graph nodes.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2018106237822 | 2018-06-15 | ||
| CN201810623782 | 2018-06-15 |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109471959A CN109471959A (en) | 2019-03-15 |
| CN109471959B true CN109471959B (en) | 2022-06-14 |
Family
ID=65678352
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811646108.2A Active CN109471959B (en) | 2018-06-15 | 2018-12-29 | Figure reasoning model-based method and system for identifying social relationship of people in image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109471959B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110111337A (en) * | 2019-04-16 | 2019-08-09 | 中山大学 | A kind of general human body analytical framework and its analytic method based on figure transfer learning |
| CN111147603A (en) * | 2019-09-30 | 2020-05-12 | 华为技术有限公司 | Method and device for networking reasoning service |
| CN114169515B (en) * | 2020-08-20 | 2023-04-28 | 四川大学 | Social relationship identification method based on high-order graph neural network |
| CN114493905B (en) * | 2020-11-13 | 2023-04-07 | 四川大学 | Social relationship identification method based on multilevel feature fusion |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104778224A (en) * | 2015-03-26 | 2015-07-15 | 南京邮电大学 | Target object social relation identification method based on video semantics |
| CN106909896A (en) * | 2017-02-17 | 2017-06-30 | 竹间智能科技（上海）有限公司 | Man-machine interactive system and method for work based on character personality and interpersonal relationships identification |
| CN106971175A (en) * | 2017-04-28 | 2017-07-21 | 深圳市唯特视科技有限公司 | It is a kind of to tangle during scene is represented the method that finds object and its relation |
| CN107016072A (en) * | 2017-03-23 | 2017-08-04 | 成都市公安科学技术研究所 | Knowledge-based inference system and method based on social networks knowledge mapping |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107735795B (en) * | 2015-07-02 | 2021-11-26 | 北京市商汤科技开发有限公司 | Method and system for social relationship identification |
-
2018
- 2018-12-29 CN CN201811646108.2A patent/CN109471959B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104778224A (en) * | 2015-03-26 | 2015-07-15 | 南京邮电大学 | Target object social relation identification method based on video semantics |
| CN106909896A (en) * | 2017-02-17 | 2017-06-30 | 竹间智能科技（上海）有限公司 | Man-machine interactive system and method for work based on character personality and interpersonal relationships identification |
| CN107016072A (en) * | 2017-03-23 | 2017-08-04 | 成都市公安科学技术研究所 | Knowledge-based inference system and method based on social networks knowledge mapping |
| CN106971175A (en) * | 2017-04-28 | 2017-07-21 | 深圳市唯特视科技有限公司 | It is a kind of to tangle during scene is represented the method that finds object and its relation |
Non-Patent Citations (4)
| Title |
|---|
| A Domain Based Approach to Social Relation Recognition;Qianru Sun 等;《2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)》;20171121;第435-444页 * |
| A social network constructed from large-scale online;Li-Jia Li 等;《ACM Transactions on Multimedia Computing, Communications, and Applications》;20150824;第12卷(第1期);第1-22页 * |
| Dual-Glance Model for Deciphering Social Relationships;Junnan Li 等;《2017 IEEE International Conference on Computer Vision (ICCV)》;20171225;第2650-2659页 * |
| 社交媒体中亲属关系的挖掘与识别;张俊康;《中国优秀博硕士学位论文全文数据库(硕士) 信息科技辑》;20180415;第I138-2765页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109471959A (en) | 2019-03-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109471959B (en) | Figure reasoning model-based method and system for identifying social relationship of people in image | |
| JP7468929B2 (en) | How to acquire geographical knowledge | |
| Zhao et al. | Image parsing with stochastic scene grammar | |
| CN109543180A (en) | A kind of text emotion analysis method based on attention mechanism | |
| CN110119449B (en) | Criminal case criminal name prediction method based on sequence-enhanced capsule network | |
| Gao et al. | CI-GNN: Building a category-instance graph for zero-shot video classification | |
| Guthier et al. | Detection and visualization of emotions in an affect-aware city | |
| CN112364168A (en) | Public opinion classification method based on multi-attribute information fusion | |
| Oluwasammi et al. | Features to text: a comprehensive survey of deep learning on semantic segmentation and image captioning | |
| Yang et al. | Prior visual relationship reasoning for visual question answering | |
| CN115809327B (en) | Real-time social network rumor detection method based on multimode fusion and topics | |
| CN114742071B (en) | Cross-language ideas object recognition analysis method based on graph neural network | |
| Kumar | Social Media Analytics for Stance Mining A Multi-Modal Approach with Weak Supervision. | |
| Zou et al. | A survey on VQA: Datasets and approaches | |
| WO2023236469A1 (en) | Video action recognition method and apparatus, electronic device, and storage medium | |
| CN110889505A (en) | Cross-media comprehensive reasoning method and system for matching image-text sequences | |
| Kong et al. | Multi-feature representation based COVID-19 risk stage evaluation with transfer learning | |
| Fang et al. | NSEP: Early fake news detection via news semantic environment perception | |
| Chen | [Retracted] Semantic Analysis of Multimodal Sports Video Based on the Support Vector Machine and Mobile Edge Computing | |
| Yang et al. | An overview of multimodal fusion learning | |
| Chae et al. | Uncertainty-based visual question answering: estimating semantic inconsistency between image and knowledge base | |
| CN115168678A (en) | Time sequence perception heterogeneous graph nerve rumor detection model | |
| CN109857884B (en) | Automatic image semantic description method | |
| Weng et al. | A survey of artificial intelligence techniques on MOOC of legal education | |
| Ide et al. | Policy decision support system in aging society based on probabilistic latent spatial semantic structure modeling |
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 |