CN116456289B - Rich media information processing method and system - Google Patents
Rich media information processing method and system Download PDFInfo
- Publication number
- CN116456289B CN116456289B CN202310697571.4A CN202310697571A CN116456289B CN 116456289 B CN116456289 B CN 116456289B CN 202310697571 A CN202310697571 A CN 202310697571A CN 116456289 B CN116456289 B CN 116456289B
- Authority
- CN
- China
- Prior art keywords
- variable
- matrix
- rich media
- output matrix
- content
- 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
- 230000010365 information processing Effects 0.000 title claims abstract description 26
- 238000003672 processing method Methods 0.000 title claims abstract description 11
- 239000011159 matrix material Substances 0.000 claims abstract description 162
- 239000013598 vector Substances 0.000 claims abstract description 100
- 238000003062 neural network model Methods 0.000 claims abstract description 20
- 238000000605 extraction Methods 0.000 claims abstract description 9
- 238000013507 mapping Methods 0.000 claims abstract description 8
- 238000012163 sequencing technique Methods 0.000 claims abstract description 6
- 238000000034 method Methods 0.000 claims description 14
- 238000004364 calculation method Methods 0.000 claims description 2
- 238000004458 analytical method Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000000306 recurrent effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/38—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
- G06F16/353—Clustering; Classification into predefined classes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/10—Text processing
- G06F40/166—Editing, e.g. inserting or deleting
- G06F40/186—Templates
-
- 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
-
- 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L51/00—User-to-user messaging in packet-switching networks, transmitted according to store-and-forward or real-time protocols, e.g. e-mail
- H04L51/07—User-to-user messaging in packet-switching networks, transmitted according to store-and-forward or real-time protocols, e.g. e-mail characterised by the inclusion of specific contents
- H04L51/10—Multimedia information
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L51/00—User-to-user messaging in packet-switching networks, transmitted according to store-and-forward or real-time protocols, e.g. e-mail
- H04L51/58—Message adaptation for wireless communication
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/12—Messaging; Mailboxes; Announcements
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Abstract
The invention relates to the technical field of short message data processing, and discloses a rich media information processing method and a rich media information processing system, wherein the rich media information processing system comprises the following components: the content extraction module is used for extracting the content of the rich media short message; a unit generation module that generates a variable unit and a non-variable unit based on the variable content and the non-variable content, respectively; the matrix generation module is used for respectively sequencing the variable unit and the non-variable unit according to the positions in the rich media short message to generate a variable unit matrix and a non-variable unit matrix; the mapping module is used for mapping the variable vector output by the RNN unit of the combined neural network model at the t time to the variable content of the t variable position of the rich media short message; the invention obtains the vector representation of each variable content and supports the vector matching of the variable content associated with the whole content of the rich media short message.
Description
Technical Field
The invention relates to the technical field of short message data processing, in particular to a rich media information processing method and a rich media information processing system.
Background
The rich media message is also called rich media short message, and compared with the traditional short message, the rich media message not only supports text, but also supports pictures, audio and video; the short message data analysis service comprises automatic garbage short message identification, big data wind control, big data marketing and the like, and only text semantic analysis is needed on the basis of traditional text short messages, but the content types of rich media messages are complex, different contents of the rich media messages are generally classified and stored, the contents of the rich media messages are decomposed, the association relationship is lost, and quick association retrieval of the short messages cannot be performed.
Disclosure of Invention
The invention provides a rich media information processing method, which solves the technical problem that the content of a rich media message is decomposed and the quick association retrieval of a short message cannot be performed in the related technology.
The invention provides a rich media information processing method, which comprises the following steps: step 101, extracting rich media short message content; extracting variable content and non-variable content; step 102, generating a variable unit based on variable content; generating a non-variable unit based on the non-variable content; step 103, respectively sequencing the variable unit and the non-variable unit according to the positions in the rich media short message to generate a variable unit matrix and a non-variable unit matrix; step 104, inputting the variable unit matrix and the non-variable unit matrix into a combined neural network model, wherein the combined neural network model comprises: the first linear layer outputs a first variable output matrix, a second variable output matrix and a third variable output matrix, and the second linear layer outputs a first non-variable output matrix; the computing of the first hidden layer includes:wherein E represents a first output matrix, D tableShowing a third variable output matrix, C representing a first variable weight matrix, the element of the j-th column of the i-th row of C being expressed as +.>；/>Wherein->Weight value representing the ith row vector of the first variable output matrix and the jth row vector of the second variable output matrix,/v>The weight value of the ith row vector of the first variable output matrix and the v row vector of the second variable output matrix is represented, and n is the number of rows of the second variable output matrix; the second hidden layer inputs the first variable output matrix, the third variable output matrix and the first non-variable output matrix, and the calculating of the second hidden layer comprises: />Wherein L represents a second output matrix, D represents a third variable output matrix, K represents a second variable weight matrix, and the element of the j-th column of the i-th row of K is represented as +.>；/>Wherein->Weight value representing the ith row vector of the first variable output matrix and the jth row vector of the first non-variable output matrix,/th row vector of the first non-variable output matrix>The weight value of the ith row vector of the first variable output matrix and the v row vector of the first non-variable output matrix is represented, and n is the number of rows of the second variable output matrix;
and after the first output matrix and the second output matrix are summed, a third output matrix is obtained and is input into a third hidden layer, an RNN (recurrent neural network) unit of the third hidden layer inputs the t-th row vector of the third output matrix at t time, and a variable vector serving as variable content of the t-th variable position of the rich media short message is output.
Further, each variable content is stored in a variable position of the rich media short message, and each non-variable content is stored in a non-variable position of the rich media short message.
Further, after the variable unit and the non-variable unit are mapped to the variable unit matrix and the non-variable unit matrix, elements in the variable unit matrix and the non-variable unit matrix that are not mapped are interpolated.
Further, row vectors in the variable cell matrix and the non-variable cell matrix represent one variable cell and one non-variable cell, respectively.
Further, the first linear layer and the second linear layer are both linearly transformed, the first variable output matrix, the second variable output matrix and the third variable output matrix are calculated to correspond to a weight matrix respectively, and the first non-variable output matrix is calculated to correspond to a weight matrix.
Further, the method comprises the steps of,wherein->An ith row vector representing a first variable output matrix,>the j-th row vector representing the second variable output matrix,>representing the telescoping parameters.
Further, the method comprises the steps of,wherein->An ith row vector representing the first variable output matrix,the j-th row vector representing the first invariant output matrix,>representing the telescoping parameters.
The invention provides a rich media information processing system, which is used for executing the rich media information processing method, and comprises the following steps:
the content extraction module is used for extracting the content of the rich media short message;
a unit generation module that generates a variable unit and a non-variable unit based on the variable content and the non-variable content, respectively;
the matrix generation module is used for respectively sequencing the variable unit and the non-variable unit according to the positions in the rich media short message to generate a variable unit matrix and a non-variable unit matrix;
a computation module for inputting the variable unit matrix and the non-variable unit matrix into the combined neural network model;
and the mapping module is used for mapping the variable vector of the output of the RNN unit of the combined neural network model at the t time to the variable content of the t variable position of the rich media short message.
Further, the rich media information processing system also comprises a retrieval module, wherein the retrieval module inputs the rich media short message with the variable unit to be retrieved into the content extraction module, and then matches the variable vector corresponding to the variable unit to be retrieved output by the combined neural network model with the variable vector in the database.
Further, the rich media information processing system further comprises a database, variable vectors stored in the database are also generated by the rich media information processing system, and variable content mapped by each variable vector is marked with a category label;
clustering is carried out from variable vectors in a database, the variable vector of a clustering center is selected as a standard vector, the variable vector corresponding to a variable unit to be searched is matched with the standard vector for marking a class label meeting the service requirement for the variable content of the standard vector, and then other variable vectors of a cluster of the standard vector are extracted for matching.
The invention has the beneficial effects that: according to the characteristic that the rich media short message is generated based on the short message template, the content of the rich media short message is extracted according to the content position, then the characteristics of expressing the whole information of the rich media short message are arranged and generated, and then the characteristics are input into a combined neural network model based on service requirements, so that vector representation of each variable content is obtained, and the variable content vectorization matching of the whole content of the related rich media short message is supported.
Drawings
FIG. 1 is a flow chart of a rich media information processing method of the present invention;
fig. 2 is a schematic block diagram of a rich media information processing system according to the present invention.
In the figure: a content extraction module 101, a unit generation module 102, a matrix generation module 103, a calculation module 104, a mapping module 105, a retrieval module 106, and a database 107.
Detailed Description
The subject matter described herein will now be discussed with reference to example embodiments. It is to be understood that these embodiments are merely discussed so that those skilled in the art may better understand and implement the subject matter described herein and that changes may be made in the function and arrangement of the elements discussed without departing from the scope of the disclosure herein. Various examples may omit, replace, or add various procedures or components as desired. In addition, features described with respect to some examples may be combined in other examples as well.
As shown in fig. 1, a rich media information processing method includes the following steps:
step 101, extracting rich media short message content;
extracting variable content and non-variable content based on the content of the rich media short message, wherein each variable content is stored in one variable position of the rich media short message, and each non-variable content is stored in one non-variable position of the rich media short message;
step 102, generating a variable unit based on variable content; generating a non-variable unit based on the non-variable content;
the variable units are in the form of vectors, and the methods for generating the variable units for different types of variable contents are different, for example, word segmentation is firstly carried out on the variable contents of the text contents, then word vectors are generated, and the generated word vectors may also need to be combined;
for example, for the variable content to be an ID, vector encoding is required for the ID according to the ID library to generate a variable unit;
for example, for image content, an image vectorization method is required to generate variable units, and in particular, convolution processing may be adopted.
For example, the video content can be decomposed into frame images and then processed;
for example, audio content may be recognized as text for processing.
Step 103, respectively sequencing the variable unit and the non-variable unit according to the positions in the rich media short message to generate a variable unit matrix and a non-variable unit matrix;
the row vectors in the variable unit matrix and the non-variable unit matrix respectively represent a variable unit or a non-variable unit;
since the matrixing process is necessary, in order to solve the problem of the difference in vector dimensions, elements that are not mapped in the variable element matrix and the non-variable element matrix are interpolated after the variable element and the non-variable element are mapped to the variable element matrix and the non-variable element matrix.
In one embodiment of the invention, 0-value interpolation is performed on elements in the variable element matrix and the non-variable element matrix that are not mapped.
Step 104, inputting the variable unit matrix and the non-variable unit matrix into a combined neural network model, wherein the combined neural network model comprises: the first linear layer outputs a first variable output matrix, a second variable output matrix and a third variable output matrix, and the second linear layer outputs a first non-variable output matrix;
the first linear layer and the second linear layer are both linearly transformed, the first variable output matrix, the second variable output matrix and the third variable output matrix are calculated to correspond to a weight matrix respectively, and the first non-variable output matrix is calculated to correspond to a weight matrix.
The computing of the first hidden layer includes:
wherein E represents a first output matrix, D represents a third variable output matrix, C represents a first variable weight matrix, and wherein the element of the j-th column of the i-th row is expressed as +.>；
Wherein->Weight value representing the ith row vector of the first variable output matrix and the jth row vector of the second variable output matrix,/v>The weight value of the ith row vector of the first variable output matrix and the v row vector of the second variable output matrix is represented, and n is the number of rows of the second variable output matrix.
Wherein->An ith row vector representing a first variable output matrix,>representing the second variable output momentThe j-th row vector of the array,>indicating a telescoping parameter, wherein the default value is 0.5;
the second hidden layer inputs the first variable output matrix, the third variable output matrix and the first non-variable output matrix, and the calculating of the second hidden layer comprises:
wherein L represents a second output matrix, D represents a third variable output matrix, K represents a second variable weight matrix, and wherein the element of the j-th column of the i-th row is represented as +.>；
Wherein->Weight value representing the ith row vector of the first variable output matrix and the jth row vector of the first non-variable output matrix,/th row vector of the first non-variable output matrix>The weight value representing the ith row vector of the first variable output matrix and the v row vector of the first non-variable output matrix, and n is the number of rows of the second variable output matrix.
Wherein->An ith row vector representing a first variable output matrix,>the j-th row vector representing the first invariant output matrix,>indicating a telescoping parameter, wherein the default value is 0.5;
after the first output matrix and the second output matrix are summed, a third output matrix is obtained and input into a third hidden layer, wherein the third hidden layer comprises an RNN (recurrent neural network) unit, and the RNN unit inputs the t-th row vector of the third output matrix at t time;
the output of the RNN unit at the time t is used as a variable vector of variable content of the t variable position of the rich media short message. This is because the t-th row vector of the third output matrix is obtained by processing the variable unit at the t-th variable position of the rich media short message.
In one embodiment of the invention, when training the combined neural network model, the outputs of the RNN units are connected to the full connection layer, and the output mapped set of classification labels of the full connection layer represent healthy content and unhealthy content, respectively.
Of course, different classification label sets can be output according to the requirements of specific short message analysis service to train the combined neural network model.
As shown in fig. 2, in one embodiment of the present invention, based on the above-mentioned method for processing rich media information, the present invention provides a rich media information processing system, which includes:
a content extraction module 101, configured to extract rich media sms content;
a unit generation module 102 that generates a variable unit and a non-variable unit based on the variable content and the non-variable content, respectively;
a matrix generating module 103, configured to respectively sequence the variable unit and the non-variable unit according to positions in the rich media sms to generate a variable unit matrix and a non-variable unit matrix;
a computation module 104 for inputting the variable element matrix and the non-variable element matrix into the combined neural network model;
a mapping module 105, configured to map a variable vector of an output of the RNN unit of the combined neural network model at time t to a variable content of a t-th variable position of the rich media sms;
the rich media information processing system further comprises a retrieval module 106, the retrieval module 106 inputs the rich media short message with the variable unit to be retrieved into the content extraction module 101, and then matches the variable vector corresponding to the variable unit to be retrieved output by the combined neural network model with the variable vector in the database 107.
In one embodiment of the present invention, the rich media information processing system further comprises a database 107, the variable vectors stored in the database 107 are also generated by the rich media information processing system, and the variable content mapped by each variable vector has been labeled with a category label that meets the business needs.
Further, clustering is performed from the variable vectors in the database 107, the variable vector in the clustering center is selected as a standard vector, and class labels meeting the service requirements are marked for the variable content of the standard vector, so that the workload of type marking can be reduced, the variable vector corresponding to the variable unit to be searched can be matched with the standard vector, then other variable vectors of the cluster of the standard vector are extracted for matching, and the matching speed can be improved.
In one embodiment of the present invention, the variable content to be searched is a short link, the short link to be searched is input to the content extraction module 101, then the variable vector corresponding to the variable unit to be searched output by the combined neural network model is matched with the variable vector in the database 107, and if the variable content corresponding to the matched variable vector is marked as the junk content, the short link to be searched is marked as the junk content. The method is applied to the processing of the junk short message identification, can accurately filter the junk short message, prevent the junk short message from being sent or the junk short message from being received, and can be used for searching for pictures or characters.
The embodiment has been described above with reference to the embodiment, but the embodiment is not limited to the above-described specific implementation, which is only illustrative and not restrictive, and many forms can be made by those of ordinary skill in the art, given the benefit of this disclosure, are within the scope of this embodiment.
Claims (10)
1. A rich media information processing method, comprising the steps of: step 101, extracting rich media short message content; extracting variable content and non-variable content; step 102, generating a variable unit based on variable content; generating a non-variable unit based on the non-variable content; step 103, respectively sequencing the variable unit and the non-variable unit according to the positions in the rich media short message to generate a variable unit matrix and a non-variable unit matrix; step 104, inputting the variable unit matrix and the non-variable unit matrix into a combined neural network model, wherein the combined neural network model comprises: the first linear layer outputs a first variable output matrix, a second variable output matrix and a third variable output matrix, and the second linear layer outputs a first non-variable output matrix; the computing of the first hidden layer includes:wherein E represents a first output matrix, D represents a third variable output matrix, C represents a first variable weight matrix, and the element of the j-th column of the i-th row of C is represented as +.>；/>Wherein->Weight value representing the ith row vector of the first variable output matrix and the jth row vector of the second variable output matrix,/v>The weight value of the ith row vector of the first variable output matrix and the v row vector of the second variable output matrix is represented, and n is the number of rows of the second variable output matrix; of a second hidden layerThe calculation includes: />Wherein L represents a second output matrix, D represents a third variable output matrix, K represents a second variable weight matrix, and the element of the j-th column of the i-th row of K is represented as +.>；Wherein->Weight value representing the ith row vector of the first variable output matrix and the jth row vector of the first non-variable output matrix,/th row vector of the first non-variable output matrix>The weight value of the ith row vector of the first variable output matrix and the v row vector of the first non-variable output matrix is represented, and n is the number of rows of the second variable output matrix;
and after the first output matrix and the second output matrix are summed, a third output matrix is obtained and is input into a third hidden layer, an RNN unit of the third hidden layer inputs the t row vector of the third output matrix at t time, and a variable vector serving as variable content of the t variable position of the rich media short message is output.
2. The method of claim 1, wherein each variable content is stored in a variable location of the rich media message and each non-variable content is stored in a non-variable location of the rich media message.
3. The method according to claim 1, wherein after the variable unit and the non-variable unit are mapped to the variable unit matrix and the non-variable unit matrix, elements in the variable unit matrix and the non-variable unit matrix that are not mapped are interpolated.
4. The method of claim 1, wherein row vectors in the variable cell matrix and the non-variable cell matrix represent a variable cell and a non-variable cell, respectively.
5. The method of claim 1, wherein the first linear layer and the second linear layer are both linearly transformed, the first variable output matrix, the second variable output matrix and the third variable output matrix are calculated to correspond to a weight matrix, and the first non-variable output matrix is calculated to correspond to a weight matrix.
6. The method for processing rich media information as set forth in claim 1, wherein,wherein->An ith row vector representing a first variable output matrix,>the j-th row vector representing the second variable output matrix,>representing the telescoping parameters.
7. The method for processing rich media information as set forth in claim 1, wherein,wherein->An ith row vector representing a first variable output matrix,>the j-th row vector representing the first invariant output matrix,>representing the telescoping parameters.
8. A rich media information processing system for executing a rich media information processing method as claimed in any one of claims 1 to 7, comprising:
the content extraction module is used for extracting the content of the rich media short message;
a unit generation module that generates a variable unit and a non-variable unit based on the variable content and the non-variable content, respectively;
the matrix generation module is used for respectively sequencing the variable unit and the non-variable unit according to the positions in the rich media short message to generate a variable unit matrix and a non-variable unit matrix;
a computation module for inputting the variable unit matrix and the non-variable unit matrix into the combined neural network model;
and the mapping module is used for mapping the variable vector of the output of the RNN unit of the combined neural network model at the t time to the variable content of the t variable position of the rich media short message.
9. The system of claim 8, further comprising a retrieval module, wherein the retrieval module inputs a rich media sms in which the variable unit to be retrieved is located into the content extraction module, and then matches a variable vector corresponding to the variable unit to be retrieved output by the combined neural network model with a variable vector in the database.
10. The rich media information processing system of claim 9 further comprising a database, wherein the variable vectors stored in the database are also generated by the rich media information processing system and the variable content mapped by each variable vector has been labeled with a class label;
clustering is carried out from variable vectors in a database, the variable vector of a clustering center is selected as a standard vector, the variable vector corresponding to a variable unit to be searched is matched with the standard vector for marking a class label meeting the service requirement for the variable content of the standard vector, and then other variable vectors of a cluster of the standard vector are extracted for matching.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310697571.4A CN116456289B (en) | 2023-06-13 | 2023-06-13 | Rich media information processing method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310697571.4A CN116456289B (en) | 2023-06-13 | 2023-06-13 | Rich media information processing method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116456289A CN116456289A (en) | 2023-07-18 |
| CN116456289B true CN116456289B (en) | 2023-08-15 |
Family
ID=87134124
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310697571.4A Active CN116456289B (en) | 2023-06-13 | 2023-06-13 | Rich media information processing method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116456289B (en) |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101281532A (en) * | 2008-05-22 | 2008-10-08 | 成都普辰瑞通通讯技术有限公司 | Expandable rich medium scene operation method |
| JP2014229142A (en) * | 2013-05-23 | 2014-12-08 | 富士通株式会社 | Program, information processing apparatus, and information processing method |
| WO2020146076A1 (en) * | 2019-01-10 | 2020-07-16 | Convida Wireless, Llc | Apparatus, system, method, and computer-readable medium for performing a message service and identity service in a 5g network |
| CN112446727A (en) * | 2019-09-04 | 2021-03-05 | 百度在线网络技术（北京）有限公司 | Advertisement triggering method, device, equipment and computer readable storage medium |
| CN114357022A (en) * | 2021-12-23 | 2022-04-15 | 北京中视广信科技有限公司 | Media content association mining method based on event relation discovery |
| WO2022161159A1 (en) * | 2021-01-28 | 2022-08-04 | 华为技术有限公司 | Message forwarding method, and system |
| CN115309563A (en) * | 2021-05-08 | 2022-11-08 | ***通信集团浙江有限公司 | 5G message generation method, device, storage medium and product |
| CN115510457A (en) * | 2021-06-04 | 2022-12-23 | ***通信集团浙江有限公司 | Data identification method, device, equipment and computer program product |
| CN115705464A (en) * | 2021-08-11 | 2023-02-17 | ***通信集团浙江有限公司 | Information processing method, device and equipment |
| CN115706940A (en) * | 2021-07-20 | 2023-02-17 | ***通信集团浙江有限公司 | Selection method and device of 5G message encryption algorithm and computing equipment |
| CN115705463A (en) * | 2021-08-11 | 2023-02-17 | ***通信集团浙江有限公司 | Information processing method, device and equipment |
| CN116206227A (en) * | 2023-04-23 | 2023-06-02 | 上海帜讯信息技术股份有限公司 | Picture examination system and method for 5G rich media information, electronic equipment and medium |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080270344A1 (en) * | 2007-04-30 | 2008-10-30 | Yurick Steven J | Rich media content search engine |
-
2023
- 2023-06-13 CN CN202310697571.4A patent/CN116456289B/en active Active
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101281532A (en) * | 2008-05-22 | 2008-10-08 | 成都普辰瑞通通讯技术有限公司 | Expandable rich medium scene operation method |
| JP2014229142A (en) * | 2013-05-23 | 2014-12-08 | 富士通株式会社 | Program, information processing apparatus, and information processing method |
| WO2020146076A1 (en) * | 2019-01-10 | 2020-07-16 | Convida Wireless, Llc | Apparatus, system, method, and computer-readable medium for performing a message service and identity service in a 5g network |
| CN112446727A (en) * | 2019-09-04 | 2021-03-05 | 百度在线网络技术（北京）有限公司 | Advertisement triggering method, device, equipment and computer readable storage medium |
| WO2022161159A1 (en) * | 2021-01-28 | 2022-08-04 | 华为技术有限公司 | Message forwarding method, and system |
| CN115309563A (en) * | 2021-05-08 | 2022-11-08 | ***通信集团浙江有限公司 | 5G message generation method, device, storage medium and product |
| CN115510457A (en) * | 2021-06-04 | 2022-12-23 | ***通信集团浙江有限公司 | Data identification method, device, equipment and computer program product |
| CN115706940A (en) * | 2021-07-20 | 2023-02-17 | ***通信集团浙江有限公司 | Selection method and device of 5G message encryption algorithm and computing equipment |
| CN115705464A (en) * | 2021-08-11 | 2023-02-17 | ***通信集团浙江有限公司 | Information processing method, device and equipment |
| CN115705463A (en) * | 2021-08-11 | 2023-02-17 | ***通信集团浙江有限公司 | Information processing method, device and equipment |
| CN114357022A (en) * | 2021-12-23 | 2022-04-15 | 北京中视广信科技有限公司 | Media content association mining method based on event relation discovery |
| CN116206227A (en) * | 2023-04-23 | 2023-06-02 | 上海帜讯信息技术股份有限公司 | Picture examination system and method for 5G rich media information, electronic equipment and medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116456289A (en) | 2023-07-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112528963A (en) | Intelligent arithmetic question reading system based on MixNet-YOLOv3 and convolutional recurrent neural network CRNN | |
| CN109685056B (en) | Method and device for acquiring document information | |
| CN108733778B (en) | Industry type identification method and device of object | |
| CN112819023B (en) | Sample set acquisition method, device, computer equipment and storage medium | |
| CN111582241A (en) | Video subtitle recognition method, device, equipment and storage medium | |
| Wilkinson et al. | Neural Ctrl-F: segmentation-free query-by-string word spotting in handwritten manuscript collections | |
| CN110210468B (en) | Character recognition method based on convolutional neural network feature fusion migration | |
| CN111914085A (en) | Text fine-grained emotion classification method, system, device and storage medium | |
| CN111523420B (en) | Header classification and header column semantic recognition method based on multi-task deep neural network | |
| CN109918506A (en) | A kind of file classification method and device | |
| CN114596566B (en) | Text recognition method and related device | |
| CN112818951A (en) | Ticket identification method | |
| CN113051914A (en) | Enterprise hidden label extraction method and device based on multi-feature dynamic portrait | |
| CN110705490B (en) | Visual emotion recognition method | |
| CN112800225B (en) | Microblog comment emotion classification method and system | |
| CN111143515B (en) | Text matching method and device | |
| CN114332473A (en) | Object detection method, object detection device, computer equipment, storage medium and program product | |
| US11494431B2 (en) | Generating accurate and natural captions for figures | |
| CN116456289B (en) | Rich media information processing method and system | |
| CN115033689B (en) | Original network Euclidean distance calculation method based on small sample text classification | |
| CN110929013A (en) | Image question-answer implementation method based on bottom-up entry and positioning information fusion | |
| CN114372128A (en) | Automatic solving method and system for rotationally symmetric geometric volume problem | |
| CN114936279A (en) | Unstructured chart data analysis method for collaborative manufacturing enterprise | |
| CN114140808A (en) | Electronic official document identification method based on domestic CPU and operating system | |
| CN110196976B (en) | Text emotional tendency classification method and device and server |
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 |