CN114022744A - Automatic illegal barrage detection method, device, system, medium and equipment - Google Patents
Automatic illegal barrage detection method, device, system, medium and equipment Download PDFInfo
- Publication number
- CN114022744A CN114022744A CN202111297497.4A CN202111297497A CN114022744A CN 114022744 A CN114022744 A CN 114022744A CN 202111297497 A CN202111297497 A CN 202111297497A CN 114022744 A CN114022744 A CN 114022744A
- Authority
- CN
- China
- Prior art keywords
- content
- video
- bullet screen
- text
- coding feature
- 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.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The application discloses a method, a device, a system, a medium and equipment for automatically detecting an illegal barrage, belonging to the field of network video. The method mainly comprises the steps of obtaining at least one piece of bullet screen content and related video content related to the bullet screen content; generating a text coding feature representative vector according to at least one piece of bullet screen content by using a text coder, and generating a video coding feature representative vector according to related video content by using a video coder; fusing the text coding characteristic representative vector and the video coding characteristic representative vector by using a multi-mode fusion device to generate a fusion coding characteristic vector; and predicting whether the at least one bullet screen content contains the violation content or not by using the result predictor according to the fusion coding feature vector to obtain a final detection result. The method and the device can automatically combine text information and video information to make correct judgment on whether the bullet screen contains illegal contents.
Description
Technical Field
The present application relates to the field of network video, and in particular, to a method, an apparatus, a system, a medium, and a device for automatically detecting an illegal barrage.
Background
The requirements for interactivity and entertainment are higher and higher when the audience watches the network video, so that a plurality of video portals are added with the function of 'barrage'. The bullet screen is to give a brief feeling and comment at any time point when a video is watched, and then the comment appears in the video and can be seen by other audiences. The barrage greatly enhances the interaction between audiences and draws the distance between the audiences and the video producer closer.
However, since the current barrage is not limited, viewers can issue barrages of any nature and content, such as barrages unrelated to video content, or even some violations, such as barrages including pornography, rumor, abuse, etc., which results in most viewers being disturbed by a small portion of such barrages when viewing video, and thus seriously affecting the viewing experience. On the other hand, although the administrator of the video portal has the authority to delete and mask the barrage, the huge amount of new videos and barrages makes it impossible for the administrator to delete all illegal barrages manually. .
Disclosure of Invention
In order to solve the problems in the prior art, the application mainly provides an automatic illegal barrage detection method, device, system, medium and equipment.
In order to achieve the above object, the present application adopts a technical solution that: an automatic violation bullet screen detection method is provided, which comprises the following steps:
acquiring at least one piece of bullet screen content and related video content related to the at least one piece of bullet screen content; generating a text coding feature representative vector according to at least one piece of bullet screen content by using a text coder, and generating a video coding feature representative vector according to related video content by using a video coder; fusing the text coding characteristic representative vector and the video coding characteristic representative vector by using a multi-mode fusion device to generate a fusion coding characteristic vector; and predicting whether the at least one bullet screen content contains the violation content or not by using the result predictor according to the fusion coding feature vector to obtain a final detection result.
Another technical scheme adopted by the application is as follows: an automatic violation bullet screen detection device is provided, which includes:
the content acquisition module is used for acquiring at least one piece of bullet screen content and related video content related to the at least one piece of bullet screen content; the representative vector generating module is used for generating a text coding characteristic representative vector according to at least one piece of bullet screen content by using a text encoder and generating a video coding characteristic representative vector according to related video content by using a video encoder; the fusion module is used for fusing the text coding characteristic representative vector and the video coding characteristic representative vector by utilizing the multi-mode fusion device to generate a fusion coding characteristic vector; and the detection module is used for predicting whether at least one bullet screen content contains the illegal content or not by using the result predictor according to the fusion coding feature vector to obtain a final detection result.
Another technical solution adopted by the application is that an automatic violation bullet screen detection system comprises a text encoder, a video encoder, a multi-module fusion device, a result predictor and the above automatic violation bullet screen detection device.
Another technical scheme adopted by the application is as follows: a computer readable storage medium storing computer instructions, wherein the computer instructions are operable to perform the automatic violation bullet screen detection method of the above-mentioned scheme.
Another technical scheme adopted by the application is as follows: a computer device comprising a processor and a memory, the memory storing computer instructions operable to perform a method of automatic violation bullet screen detection in accordance with one of the above aspects.
The technical scheme of the application can reach the beneficial effects that: the application designs an automatic illegal barrage detection method, device, system, medium and equipment. According to the method and the device, the text features and the video features are fused, the predictor is used for predicting according to the fused features, and the text information and the video information can be automatically combined to make correct judgment on whether the bullet screen contains illegal contents, so that operations such as selecting, retaining, shielding or filtering are immediately and automatically performed on the bullet screen, and the auditing time is greatly reduced.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present application, and for those skilled in the art, other drawings can be obtained according to these drawings without inventive exercise.
Fig. 1 is a schematic flow chart of an embodiment of an automatic violation bullet screen detection method according to the present application;
fig. 2 is a schematic flowchart of an embodiment of an automatic illegal barrage detection method according to the present application;
fig. 3 is a schematic flowchart of an embodiment of an automatic violation bullet screen detection method according to the present application;
fig. 4 is a schematic diagram of an embodiment of an automatic violation bullet screen detection device according to the present application;
fig. 5 is a schematic diagram of an embodiment of an automatic violation bullet screen detection device according to the present application.
With the foregoing drawings in mind, certain embodiments of the disclosure have been shown and described in more detail below. These drawings and written description are not intended to limit the scope of the disclosed concepts in any way, but rather to illustrate the concepts of the disclosure to those skilled in the art by reference to specific embodiments.
Detailed Description
The following detailed description of the preferred embodiments of the present application, taken in conjunction with the accompanying drawings, will provide those skilled in the art with a better understanding of the advantages and features of the present application, and will make the scope of the present application more clear and definite.
It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
With the development of internet technology, the requirements of audiences on interactivity and entertainment are higher and higher when watching network videos, so that a plurality of video portals are added with a 'barrage' function. The bullet screen is to give a brief feeling and comment at any time point when a video is watched, and then the comment appears in the video and can be seen by other audiences. The barrage greatly enhances the interaction between audiences and draws the distance between the audiences and the video producer closer.
However, since the current barrage is not limited, viewers can issue barrages of any nature and content, such as barrages unrelated to video content, or even some violations, such as barrages including pornography, rumor, abuse, etc., which results in most viewers being disturbed by a small portion of such barrages when viewing video, and thus seriously affecting the viewing experience. On the other hand, although the administrator of the video portal has the authority to delete, mask, barrage, the amount of hundreds of thousands of new videos per day, such as marine barrage, makes it impossible for the administrator to manually delete all violating barrages.
When judging whether the bullet screen contains the illegal content, if the bullet screen is only judged according to the content of the bullet screen, whether the bullet screen contains the illegal content cannot be accurately distinguished. For example, if the video content is a pig, the bullet screen content does not include the content related to the suspected abuse; but are suspected of containing abuse violations if the video content is human.
The application designs an automatic illegal barrage detection method, device, system, medium and equipment based on multi-mode information, and after a user sends a barrage, the user can immediately and automatically judge whether the barrage is illegal, so that operations such as reserving, shielding or filtering are selected, and the manual examination time is greatly reduced.
The technical solution of the present application is described in detail below with specific embodiments and with reference to the accompanying drawings. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments.
Fig. 1 shows a specific embodiment of the automatic illegal barrage detection method according to the present application.
In this specific embodiment, the method for automatically detecting an illegal barrage mainly includes: s101, acquiring at least one piece of barrage content and related video content related to the at least one piece of barrage content; s102, generating a text coding characteristic representative vector according to at least one bullet screen content by using a text coder, and generating a video coding characteristic representative vector according to related video content by using a video coder; s103, fusing the text coding characteristic representative vector and the video coding characteristic representative vector by using a multi-mode fusion device to generate a fusion coding characteristic vector; and S104, predicting whether the at least one bullet screen content contains the violation content or not by using the result predictor according to the fusion coding feature vector to obtain a final detection result.
By fusing the text features and the video features and predicting according to the fused features by using the predictor, correct judgment can be made on whether the bullet screen contains illegal contents or not by automatically combining the text information and the video information, so that operations such as selective retention, shielding or filtering and the like can be performed on the bullet screen immediately, and the auditing time is greatly reduced.
The process of acquiring at least one bullet screen content and related video content related to the at least one bullet screen content, which is shown in the process S101 in fig. 1, can facilitate further processing and merging according to the bullet screen content and the related video content, and determine whether the bullet screen content contains violation information.
In an embodiment of the present application, one piece of barrage content is acquired at a time, so as to further process and merge the piece of barrage content and the related video content and determine whether the piece of barrage content contains violation information.
In a specific embodiment of the present application, a plurality of bullet screen contents are acquired at a time, so as to further process and merge the plurality of bullet screen contents and the related video contents and determine whether the bullet screen contents contain violation information.
In a specific embodiment of the present application, the related video content is a video content within a predetermined time period before and after a time point of the at least one barrage content appearing.
In a specific embodiment of the present application, the content in the predetermined time period before and after the time point of the bullet-screen content appearance is the video content within 5 seconds before and after the time point of the bullet-screen content appearance, which is 10 seconds in total.
In a specific example of the present application, as shown in fig. 3, several words of "this teacher's real commander" of the bullet screen X appear, find the time point T of its appearance, take 5 seconds before and after the time point T, and use the video content within these ten seconds as the related video content V of the bullet screen X.
The process S102 in fig. 1 illustrates a process of generating a text coding feature representative vector according to at least one bullet screen content by using a text encoder, and generating a video coding feature representative vector according to a related video content by using a video encoder, where a fusion vector can be obtained according to the text coding feature representative vector and the video coding feature representative vector, so as to further facilitate a determination of whether the current bullet screen contains violation information according to the fusion vector.
In a specific embodiment of the present application, the process of generating the text coding feature representative vector according to the at least one barrage content by using the text encoder and generating the video coding feature representative vector according to the related video content by using the video encoder includes generating a corresponding text coding feature vector according to each word of the at least one barrage content by using the text encoder as the text coding feature representative vector and generating a corresponding video coding feature vector according to each second of the video content by using the video encoder as the video coding feature representative vector.
In one embodiment of the present application, as shown in fig. 3, the barrage X and the video content V are respectively fed into a text encoder and a video encoder, and respective encoding feature vectors are obtained. The coded feature vectors are a plurality of vectors, the number of the text coded feature vectors is equal to the length of a bullet screen X, in the figure, the bullet screen is 'the teacher really is commander', the length is 6, the text coded feature vectors are totally 6, and each feature represents a word; and a total of 10 video coding feature vectors are provided, corresponding to each of the 10 seconds T-5 to T + 5.
In an embodiment of the application, the process of generating the text coding feature representative vector according to the at least one barrage content by using the text encoder and generating the video coding feature representative vector according to the related video content by using the video encoder includes generating a corresponding text coding feature vector according to each word of the at least one barrage content by using the text encoder, generating a corresponding video coding feature vector according to each second of video content of the related video content by using the video encoder, performing maximum pooling on all the text coding feature vectors to obtain the text coding feature representative vector, and performing maximum pooling on all the video coding feature vectors to obtain the video coding feature representative vector. The text coding feature vector is subjected to maximum pooling to obtain a single vector, and the video coding feature vector is subjected to maximum pooling to obtain a single vector, so that the calculation is simplified when the final violation judgment is carried out conveniently.
In a specific example of the present application, the maximum pooling process comprises mapping 6 text-coding features intoQuantity "x1,x2,…x6"and video coding feature vector" v1，v2…v10' respectively carrying out maximum pooling to obtain text coding feature representative vectors
And forming a video coding feature representative vector
The process S103 in fig. 1 represents that the multi-modal fuser is used to fuse the text coding feature representative vector and the video coding feature representative vector to generate a fused coding feature vector, so that the text information and the video information can be automatically fused, and it is beneficial to subsequently further determine whether the current bullet screen contains the violation content according to the fused coding feature vector.
In an embodiment of the application, the process of fusing the text coding feature representative vector and the video coding feature representative vector by using the multi-modal fuser to generate the fused coding feature vector includes sending the text coding feature representative vector and the video coding feature representative vector to the multi-modal fuser together, and allowing them to interact sufficiently to obtain the fused coding feature vector.
In the specific example shown in FIG. 3, the process of sending the video-encoded feature representative vectors of the text-encoded feature representative vectors into the multi-modal fusion device to be fully interacted with each other to obtain the fused encoded feature vectors includes sending the text-encoded feature representative vectors into the multi-modal fusion device
And become videoCoding feature representative vector
Sending the fusion products into a multi-modal fusion device for fusion by the following formula:
a fused encoded feature vector may be obtained.
As indicated by the process S104 in fig. 1, the result predictor is used to predict whether at least one bullet screen content includes the illegal content according to the fusion coding feature vector to obtain a final detection result, and it is possible to correctly determine whether the current bullet screen content includes the illegal content according to the fusion coding feature vector combining the text information and the video information, so that operations such as selecting, retaining, shielding, or filtering are performed on the bullet screen in real time, and the auditing time is greatly reduced.
In a specific example of the present application, the result predictor is a result predictor in the prior art, which is subjected to linear transformation, a vector w is point-multiplied by a vector z obtained in one step to output a scalar value q, then the scalar value is transformed to the interval range of (0,1) by using a sigmoid function, and if the obtained number is greater than 0.5, the output result is considered to be "yes"; if less than 0.5, the output result is considered as "no". The process is as follows:
the binary value "yes" or "no" is then output. If yes, it indicates that the at least one subtitle content includes violation information; if the output is 'no', the fact that the at least one subtitle content does not include violation information is indicated.
In a specific example of the present application, as shown in fig. 2, the method for automatically detecting an illegal barrage further includes a process S201 of monitoring barrage content in real time, so as to obtain at least one piece of barrage content and related video content at a first time.
When the video is played, the barrage is monitored in real time, so that the subtitles and related video content can be acquired at the first time when the subtitles appear, illegal content detection is further performed, operations such as selecting, retaining, shielding or filtering and the like on the barrage in real time are facilitated, and the auditing time is greatly reduced.
Fig. 4 shows an embodiment of an automatic illegal barrage detection device according to the present application.
In this embodiment, the automatic illegal barrage detection apparatus mainly includes a content obtaining module 401, configured to obtain at least one barrage content and related video content related to the at least one barrage content; a representative vector generating module 402, configured to generate a representative vector of text coding features according to at least one piece of bullet screen content by using a text encoder, and generate a representative vector of video coding features according to related video content by using a video encoder; the fusion module 403 is configured to fuse the text coding feature representative vector and the video coding feature representative vector by using a multi-modal fusion device to generate a fusion coding feature vector; and the detection module 404 is configured to predict whether the at least one bullet screen content includes the violation content according to the fusion coding feature vector by using the result predictor, so as to obtain a final detection result.
The text features and the video features are fused, the predictor is used for predicting according to the fused features, and correct judgment can be made on whether the bullet screen contains illegal contents or not by automatically combining the text information and the video information, so that operations such as selective retention, shielding or filtering and the like can be performed on the bullet screen immediately, and the auditing time is greatly reduced.
The content obtaining module 401 is configured to obtain at least one bullet screen content and related video content related to the at least one bullet screen content, and can facilitate further processing and fusing according to the bullet screen content and the related video content, and determine whether the bullet screen content includes violation information.
In a specific embodiment of the present application, the related video content is a video content within a predetermined time period before and after a time point of the at least one barrage content appearing.
The representative vector generating module 402 is configured to generate a text coding feature representative vector according to at least one bullet screen content by using a text encoder, and generate a video coding feature representative vector according to related video content by using a video encoder, so that a fusion vector can be obtained according to the text coding feature representative vector and the video coding feature representative vector, and further, whether the current bullet screen contains violation information or not can be judged according to the fusion vector.
In one embodiment of the present application, the above-mentioned representative vector generating module 402 includes an encoding feature vector generating sub-module, as shown in block 503 of fig. 5, capable of generating a corresponding text encoding feature vector according to each word of at least one piece of bullet screen content by using a text encoder, and generating a corresponding video encoding feature vector according to each second of video content of the related video content by using a video encoder.
In a specific example of the present application, the representative vector generating module 402 uses the text encoding feature vector as a text encoding feature representative vector, and uses the video encoding feature vector as a video encoding feature representative vector.
In a specific embodiment of the present application, the representative vector generating module 402 includes a max-pooling sub-module, as shown in a module 503 of fig. 5, configured to max-pool all the text-encoded feature vectors to obtain text-encoded feature representative vectors, and max-pool all the video-encoded feature vectors to obtain video-encoded feature representative vectors.
In one embodiment of the present application, the representative vector generation module 402 includes the coded feature vector generation sub-module and a max-pooling sub-module.
The fusion module 403 is configured to fuse the text coding feature representative vector and the video coding feature representative vector by using the multi-modal fusion device to generate a fusion coding feature vector, so that the text information and the video information can be automatically fused together, and it is further facilitated to subsequently determine whether the current bullet screen includes the violation content according to the fusion coding feature vector.
In one embodiment of the present application, the fusion module 403 sends the video coding feature representative vectors of the text coding feature representative vectors to the multi-modal fusion device, so that the video coding feature representative vectors are fully interacted to obtain the fusion coding feature vectors.
And the detection module 404 is configured to predict whether the at least one bullet screen content includes the violation content according to the fusion coding feature vector by using the result predictor, so as to obtain a final detection result. Whether the current bullet screen content contains illegal content or not can be judged correctly according to the fusion coding feature vector combining the text information and the video information, so that operations such as selective retention, shielding or filtering and the like are carried out on the bullet screen immediately, and the auditing time is greatly reduced.
In one embodiment of the present application, the detection module 404 can send the fused encoded feature vector to the result predictor, and output a binary value, yes or no, as the final prediction result.
In an embodiment of the present application, the automatic illegal barrage detection apparatus further includes a detection module, as shown in a module 501 of fig. 5, configured to monitor barrage content in real time, so as to obtain the barrage content and related video content at the first time.
When the video is played, the barrage is monitored in real time, so that the subtitles and related video content can be acquired at the first time when the subtitles appear, illegal content detection is further performed, operations such as selecting, retaining, shielding or filtering and the like on the barrage in real time are facilitated, and the auditing time is greatly reduced.
In an embodiment of the present application, the functional modules in an automatic violation bullet screen detection apparatus of the present application may be directly in hardware, in a software module executed by a processor, or in a combination of the two.
In another embodiment of the present application, an automatic illegal barrage detection system includes a text encoder, a video encoder, a multi-module fusion device, a result predictor, and the above automatic illegal barrage detection device.
A software module may reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor such the processor can read information from, and write information to, the storage medium.
The Processor may be a Central Processing Unit (CPU), other general-purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA), other Programmable logic devices, discrete Gate or transistor logic, discrete hardware components, or any combination thereof. A general purpose processor may be a microprocessor, but in the alternative, the processor may be any conventional processor, controller, microcontroller, or state machine. A processor may also be implemented as a combination of computing devices, e.g., a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a DSP core, or any other such configuration. In the alternative, the storage medium may be integral to the processor. The processor and the storage medium may reside in an ASIC. The ASIC may reside in a user terminal. In the alternative, the processor and the storage medium may reside as discrete components in a user terminal.
In another embodiment of the present application, a computer-readable storage medium stores computer instructions, which are operated by the automatic violation bullet screen detection method in the above-mentioned scheme.
In another embodiment of the present application, a computer device includes a processor and a memory, the memory storing computer instructions, the computer instructions being operative to perform the automatic violation bullet screen detection method in the above scheme.
In the several embodiments provided in the present application, it should be understood that the disclosed apparatus and method may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the units is only one logical division, and other divisions may be realized in practice, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
The above description is only an example of the present application and is not intended to limit the scope of the present application, and all equivalent structural changes made by using the contents of the specification and the drawings, which are directly or indirectly applied to other related technical fields, are included in the scope of the present application.
Claims (10)
1. An automatic illegal barrage detection method is characterized by comprising the following steps,
acquiring at least one piece of barrage content and related video content related to the at least one piece of barrage content;
generating a text coding feature representative vector according to the at least one bullet screen content by using a text encoder, and generating a video coding feature representative vector according to the related video content by using a video encoder;
fusing the text coding feature representative vector and the video coding feature representative vector by utilizing a multi-mode fusion device to generate a fusion coding feature vector; and the number of the first and second groups,
and predicting whether the at least one bullet screen content contains violation content or not by using a result predictor according to the fusion coding feature vector to obtain a final detection result.
2. The automatic illegal barrage detection method according to claim 1, characterized in that the related video content comprises video content within a predetermined time period before and after a time point of occurrence of the at least one barrage content.
3. The automatic illegal barrage detection method according to claim 1, wherein the process of generating a text-encoded feature representative vector from the at least one barrage content using a text encoder comprises,
generating a corresponding text coding feature vector according to each word of the bullet screen content by using the text encoder, and performing maximum pooling on all the text coding feature vectors to obtain the text coding feature representative vector;
the process of generating a video coding feature representative vector from the associated video content using a video encoder comprises,
and generating a corresponding video coding feature vector according to each second of video content of the related video content by using the video encoder, and performing maximum pooling on all the video coding feature vectors to obtain the video coding feature representative vector.
4. The automatic violation bullet screen detection method of claim 1, further comprising,
and monitoring the bullet screen content in real time so as to acquire the at least one piece of bullet screen content and the related video content at the first time.
5. An automatic illegal barrage detection device is characterized by comprising,
the content acquisition module is used for acquiring at least one piece of bullet screen content and related video content related to the bullet screen content;
the representative vector generating module is used for generating a text coding characteristic representative vector according to the at least one bullet screen content by using a text encoder and generating a video coding characteristic representative vector according to the related video content by using a video encoder;
the fusion module is used for fusing the text coding characteristic representative vector and the video coding characteristic representative vector by utilizing a multi-mode fusion device to generate a fusion coding characteristic vector; and the number of the first and second groups,
and the detection module is used for predicting whether the at least one bullet screen content contains the illegal content or not by using a result predictor according to the fusion coding feature vector to obtain a final detection result.
6. The automatic violation bullet screen detection device of claim 5, further comprising,
and the monitoring module is used for monitoring the bullet screen content in real time so as to obtain the bullet screen content and the related video content at the first time.
7. The automatic illegal barrage detection device according to claim 5, wherein the representative vector generation module comprises a coding feature vector generation submodule and a maximum pooling submodule;
the coding feature vector generation submodule is used for generating a corresponding text coding feature vector according to each word of the at least one bullet screen content by using the text encoder and generating a corresponding video coding feature vector according to each second of video content of the related video content by using the video encoder;
the maximum pooling sub-module is used for performing maximum pooling on all the text coding feature vectors to obtain the text coding feature representative vector, and performing maximum pooling on all the video coding feature vectors to obtain the video coding feature representative vector.
8. An automatic violation bullet screen detection system comprising a text encoder, a video encoder, a multi-module fuser, a result predictor and an automatic violation bullet screen detection apparatus according to any of claims 5-7.
9. A computer readable storage medium storing computer instructions, wherein the computer instructions are operable to perform the automatic violation bullet screen detection method of any of claims 1-4.
10. A computer device comprising a processor and a memory, the memory storing computer instructions, wherein the processor operates the computer instructions to perform the automatic violation bullet screen detection of any of claims 1-4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111297497.4A CN114022744A (en) | 2021-11-04 | 2021-11-04 | Automatic illegal barrage detection method, device, system, medium and equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111297497.4A CN114022744A (en) | 2021-11-04 | 2021-11-04 | Automatic illegal barrage detection method, device, system, medium and equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114022744A true CN114022744A (en) | 2022-02-08 |
Family
ID=80060748
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111297497.4A Pending CN114022744A (en) | 2021-11-04 | 2021-11-04 | Automatic illegal barrage detection method, device, system, medium and equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114022744A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117370934A (en) * | 2023-12-04 | 2024-01-09 | 环球数科集团有限公司 | Multi-mode data enhancement method of sensitive information discovery model |
-
2021
- 2021-11-04 CN CN202111297497.4A patent/CN114022744A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117370934A (en) * | 2023-12-04 | 2024-01-09 | 环球数科集团有限公司 | Multi-mode data enhancement method of sensitive information discovery model |
| CN117370934B (en) * | 2023-12-04 | 2024-03-22 | 环球数科集团有限公司 | Multi-mode data enhancement method of sensitive information discovery model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20210397651A1 (en) | Estimating social interest in time-based media | |
| CN108683924B (en) | Video processing method and device | |
| CN110072142B (en) | Video description generation method and device, video playing method and device and storage medium | |
| TWI590650B (en) | Object tracking in encoded video streams | |
| CN110012302A (en) | A kind of network direct broadcasting monitoring method and device, data processing method | |
| Shi et al. | A fast and robust key frame extraction method for video copyright protection | |
| US9609338B2 (en) | Layered video encoding and decoding | |
| WO2014089145A2 (en) | Methods and systems for displaying contextually relevant information regarding a media asset | |
| KR20140043406A (en) | Highly scalable audience measurement system with client event pre-processing | |
| CN110287375B (en) | Method and device for determining video tag and server | |
| CN114022744A (en) | Automatic illegal barrage detection method, device, system, medium and equipment | |
| CN115150384B (en) | Point cloud encoding and decoding method and device, computer readable medium and electronic equipment | |
| CN112015736B (en) | Multi-functional recommendation method and device based on Spark Mllib | |
| CN112073757A (en) | Emotion fluctuation index acquisition method, emotion fluctuation index display method and multimedia content production method | |
| CN107948755B (en) | Video content recommendation method and system combining user watching duration | |
| CN112437289B (en) | Switching time delay obtaining method | |
| CN108881950B (en) | Video processing method and device | |
| US9357178B1 (en) | Video-revenue prediction tool | |
| CN112347274B (en) | Recommendation method and system of network media assets, computer equipment and storage medium | |
| CN110505501A (en) | Information processing method, electronic device, and computer-readable storage medium | |
| CN117278780B (en) | Video encoding and decoding method, device, equipment and storage medium | |
| CN117061815A (en) | Video processing method, video processing device, computer readable medium and electronic equipment | |
| CN117939252A (en) | Data processing method, device, equipment and storage medium | |
| CN108271072B (en) | Evaluation method and device of electronic navigation menu | |
| CN115391663A (en) | Video recommendation method and device, computer equipment and storage medium |
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 |