Live broadcast platform content auditing intelligent evaluation system
By introducing data crawling, preprocessing, analysis, calculation, storage and evaluation modules into the live broadcast platform content audit intelligent evaluation system, the problem of lack of scoring system is solved, automatic quality inspection and visual report generation is realized, and operators are reduced work burden and manual review costs.
Patent Information
- Application Number
- CN202410943178.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-15
- Publication Date
- 2024-11-08
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The live broadcast platform's content review intelligent evaluation system lacks a scoring system for the live broadcast room and cannot output quality inspection reports from management perspectives, resulting in operators needing to frequently view and count past live broadcast records, which is inconvenient to operate.
An intelligent evaluation system for content review of live broadcast platforms is designed, including data crawling, preprocessing, analysis, calculation, storage, evaluation and feedback modules. By automatically crawling and analyzing live broadcast content, quality inspection and evaluation are carried out to generate visual quality. Check reports and backup and manage data in the cloud, covering multiple audit dimensions and reducing manual audit costs.
It realizes automatic analysis and identification of live broadcast content, generates visual quality inspection results, facilitates management personnel to view and evaluate, reduces the burden of frequent statistics, and reduces platform risks and manual review costs through real-time audits.
Smart Images

Figure CN118917683A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of intelligent evaluation system for content review of live broadcast platforms, and in particular to an intelligent evaluation system for content review of live broadcast platforms. Background Art
[0002] The intelligent evaluation system for content review on live streaming platforms is an intelligent tool for evaluating and reviewing content on live streaming platforms. It uses artificial intelligence technology and algorithms to automatically detect and analyze live streaming content to determine whether it complies with the platform's content specifications and legal and regulatory requirements. The system can evaluate various types of content, including but not limited to text, images, audio and video. It can automatically detect and filter out content that violates platform regulations or is inappropriate, such as pornography, violence, malicious attacks, illegal or prohibited content, etc. However, when using the intelligent evaluation system for content review on live streaming platforms, there is usually a lack of a scoring system for live streaming rooms, and quality inspection reports from a management perspective cannot be output. When operators need to view live streaming content indicators and evaluations, they need to frequently view and count past live streaming records, which is inconvenient. Summary of the invention
[0003] In order to overcome the problem that when using the intelligent evaluation system for content review on the live broadcast platform, there is usually a lack of a scoring system for the live broadcast room and the inability to output quality inspection reports from a management perspective. When operators need to view the live broadcast content indicators and evaluations, they need to frequently view and count past live broadcast records, which is quite inconvenient.
[0004] The technical solution of the present invention is: a live broadcast platform content review intelligent evaluation system, including:
[0005] Data capture module: used to obtain video and audio data from live streams or uploaded media;
[0006] Data preprocessing module: used to perform initial processing on the data captured by the data capture module;
[0007] Data analysis module: used to extract and analyze the initially processed video and audio files;
[0008] Data calculation module: used to perform quality inspection on the extracted data and determine whether it is compliant;
[0009] Data storage module: used to back up and store the extracted and quality-inspected data;
[0010] Data evaluation module: used to evaluate the overall live broadcast data indicators and live broadcast content;
[0011] Data feedback module: used to allow users to report and provide feedback on content, and to conduct secondary confirmation of uncertain content.
[0012] Preferably, the live content and uploaded content of the live broadcast platform are extracted through the data capture module, the captured data is preliminarily processed through the data preprocessing module to remove interference information, the preprocessed media feature content is analyzed and classified through the data analysis module, the extracted live data content is quality inspected and judged through the data calculation module, the extracted and quality inspected data are backed up in the cloud through the data storage module to prevent loss, the live broadcast situation and content are evaluated and scored through the data evaluation module, and the data feedback module facilitates the collection of user information feedback.
[0013] Preferably, the data capture module is used to scan and extract data information of the user's live broadcast content or uploaded content. The data capture module can automatically obtain the content and related information of different live broadcast rooms in the live broadcast platform, and transmit the captured content and related information to the data preprocessing module.
[0014] Preferably, the data capture module includes the following method when capturing live broadcast information and content:
[0015] A11: When capturing text data, directly capture the text data that can be directly collected from the bullet screen, comments, etc. on the live broadcast platform;
[0016] A12: When capturing image data information, perform text extraction or intelligent recognition on live video streams, expressions, images and other related graphics and texts;
[0017] A13: When extracting information from animations and interactive content, capture the information data in the video and audio by recording the screen.
[0018] Preferably, the data preprocessing module is used to perform preliminary processing on the live broadcast information and content captured by the data capture module, remove noise, inconsistent formats or other messy parts in the captured data, which can improve the accuracy and efficiency of subsequent data reading, while reducing the impact of interference parts on the sensitivity of the data analysis module, and transmit the preprocessed data to the data analysis module.
[0019] Preferably, the data analysis module includes the following steps when analyzing the preprocessed data:
[0020] S11: Obtain pre-processed live video information and content;
[0021] S12: extracting the audio and text information in the video file;
[0022] S13: Convert the recording file into a text file;
[0023] S14: Analyze the keyword frequency and time.
[0024] Preferably, when the data calculation module performs quality inspection on the live broadcast content and information, the following steps are included:
[0025] S21: Create quality inspection rules, which include inspection scope, quality inspection regulations, time budget and time plan;
[0026] S22: Perform quality inspection calculations to conduct multiple audit dimensions on the video stream, audio stream, bullet screen\comments, etc. in the live content;
[0027] S23: After the quality inspection is completed, the quality inspection results and quality inspection report are obtained;
[0028] Preferably, through the data calculation module, the live broadcast content can be randomly inspected, and the video stream, audio stream, barrage / comments and other content on the live broadcast platform can be reviewed in real time, covering multiple review dimensions such as pornography, violence and terrorism, political sensitivity, advertising, prohibited items, etc., greatly reducing platform risks and manual review costs.
[0029] Preferably, the data storage module includes an information resource management submodule and a cloud storage terminal, which integrates the captured information and the quality inspection completion data, manages and stores the integrated data, and stores and backs up the data to the cloud.
[0030] Preferably, when the data storage module backs up and stores data, the following method is included:
[0031] A21: Create a master document based on the corresponding date on the cloud platform;
[0032] A22: Create classification sub-documents in the main document according to the classification information;
[0033] A23: Compress and encrypt the live broadcast data and content for backup;
[0034] A24: The compressed and encrypted live broadcast data and content are placed in a sub-document.
[0035] Preferably, the data evaluation module includes the following indicators when evaluating the overall live data indicators and content:
[0036] A31: Data indicators, including the number of viewers in the live broadcast room, thousands of leads, live broadcast duration and other data indicators;
[0037] A32: Keyword frequency, the number of times a keyword appears per hour in a live broadcast;
[0038] A33: Number of keywords, the magnitude of the accumulated number of keywords in a live broadcast;
[0039] A34: Oral speaking speed, counting the average number of oral speaking times per minute;
[0040] A35: Silence. This refers to a period of long-term cessation of content output during live broadcast. A period of silence lasting up to three minutes is considered as one period of continuous silence.
[0041] Preferably, the live broadcast content is automatically analyzed and identified through the data evaluation module, and the data and content during the live broadcast are evaluated and scored. This can produce visual project comprehensive quality inspection results, as well as quality inspection result data management in multiple dimensions such as the live broadcast room, dealer stores, and regions, avoiding the need for staff to perform frequent statistical reviews when evaluating and viewing the live broadcast data.
[0042] Preferably, the data feedback module is used to collect and analyze evaluation feedback information of the overall live broadcast data, collect data on the accuracy and reliability of the content review system through user reports, complaints and other feedback methods, analyze user feedback information, and optimize the system's performance in identifying problems and misjudging problems.
[0043] Beneficial effects of the present invention:
[0044] 1. Compared with the traditional live broadcast platform content review intelligent evaluation system, it usually lacks a scoring system for live broadcast rooms and cannot output quality inspection reports from a management perspective. When operators need to view live broadcast content indicators and evaluations, they need to frequently view and count past live broadcast records, which is inconvenient. The live broadcast platform content review intelligent evaluation system outputs visual project comprehensive quality inspection and quality inspection result data management in multiple dimensions of live broadcast rooms, distribution stores, and regions, making it convenient for staff to view and evaluate live broadcast content and data;
[0045] 2. The data evaluation module can automatically analyze and identify live broadcast content, evaluate and score the data and content during live broadcast, and produce visual project comprehensive quality inspection results, as well as quality inspection result data management in multiple dimensions such as live broadcast room, dealer stores, and regions, so as to avoid the need for staff to conduct frequent statistical checks when evaluating and viewing live broadcast data;
[0046] 3. Through the data calculation module, the live broadcast content can be randomly inspected, and the video stream, audio stream, barrage / comments and other content on the live broadcast platform can be reviewed in real time, covering multiple review dimensions such as pornography, violence and terrorism, political sensitivity, advertising, prohibited items, etc., greatly reducing platform risks and manual review costs. BRIEF DESCRIPTION OF THE DRAWINGS
[0047] Figure 1 What is shown is a schematic diagram of the workflow of the data analysis module of the present invention;
[0048] Figure 2 What is shown is a schematic diagram of the workflow of the data calculation module of the present invention; DETAILED DESCRIPTION
[0049] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.
[0050] See also Figure 1 The present invention provides an embodiment: a live broadcast platform content review intelligent evaluation system, comprising:
[0051] Data capture module: used to obtain video and audio data from live streams or uploaded media;
[0052] Data preprocessing module: used to perform initial processing on the data captured by the data capture module;
[0053] Data analysis module: used to extract and analyze the initially processed video and audio files;
[0054] Data calculation module: used to perform quality inspection on the extracted data and determine whether it is compliant;
[0055] Data storage module: used to back up and store the extracted and quality-inspected data;
[0056] Data evaluation module: used to evaluate the overall live broadcast data indicators and live broadcast content;
[0057] Data feedback module: used to allow users to report and provide feedback on content, and to conduct secondary confirmation of uncertain content.
[0058] See also Figure 2 In this embodiment, as a preference, the data capture module is used to scan and extract data information of the user's live broadcast content or uploaded content. The data capture module can automatically obtain the content and related information of different live broadcast rooms in the live broadcast platform, and transmit the captured content and related information to the data preprocessing module.
[0059] Preferably, the data capture module includes the following method when capturing live broadcast information and content:
[0060] A11: When capturing text data, directly capture the text data that can be directly collected from the bullet screen, comments, etc. on the live broadcast platform;
[0061] A12: When capturing image data information, perform text extraction or intelligent recognition on live video streams, expressions, images and other related graphics and texts;
[0062] A13: When extracting information from animations and interactive content, capture the information data in the video and audio by recording the screen.
[0063] Preferably, the data preprocessing module is used to perform preliminary processing on the live broadcast information and content captured by the data capture module, remove noise, inconsistent formats or other messy parts in the captured data, which can improve the accuracy and efficiency of subsequent data reading, while reducing the impact of interference parts on the sensitivity of the data analysis module, and transmit the preprocessed data to the data analysis module.
[0064] Preferably, the data analysis module includes the following steps when analyzing the preprocessed data:
[0065] S11: Obtain pre-processed live video information and content;
[0066] S12: extracting the audio and text information in the video file;
[0067] S13: Convert the recording file into a text file;
[0068] S14: Analyze the keyword frequency and time.
[0069] Preferably, when the data calculation module performs quality inspection on the live broadcast content and information, the following steps are included:
[0070] S21: Create quality inspection rules, which include inspection scope, quality inspection regulations, time budget and time plan;
[0071] S22: Perform quality inspection calculations to conduct multiple audit dimensions on the video stream, audio stream, bullet screen\comments, etc. in the live content;
[0072] S23: After the quality inspection is completed, the quality inspection results and quality inspection report are obtained;
[0073] Preferably, through the data calculation module, the live broadcast content can be randomly inspected, and the video stream, audio stream, barrage / comments and other content on the live broadcast platform can be reviewed in real time, covering multiple review dimensions such as pornography, violence and terrorism, political sensitivity, advertising, prohibited items, etc., greatly reducing platform risks and manual review costs.
[0074] Preferably, the data storage module includes an information resource management submodule and a cloud storage terminal, which integrates the captured information and the quality inspection completion data, manages and stores the integrated data, and stores and backs up the data to the cloud.
[0075] Preferably, when the data storage module backs up and stores data, the following method is included:
[0076] A21: Create a master document based on the corresponding date on the cloud platform;
[0077] A22: Create classification sub-documents in the main document according to the classification information;
[0078] A23: Compress and encrypt the live broadcast data and content for backup;
[0079] A24: The compressed and encrypted live broadcast data and content are placed in a sub-document.
[0080] Preferably, the data evaluation module includes the following indicators when evaluating the overall live data indicators and content:
[0081] A31: Data indicators, including the number of viewers in the live broadcast room, thousands of leads, live broadcast duration and other data indicators;
[0082] A32: Keyword frequency, the number of times a keyword appears per hour in a live broadcast;
[0083] A33: Number of keywords, the magnitude of the accumulated number of keywords in a live broadcast;
[0084] A34: Oral speaking speed, counting the average number of oral speaking times per minute;
[0085] A35: Silence. This refers to a period of long-term cessation of content output during live broadcast. A period of silence lasting up to three minutes is considered as one period of continuous silence.
[0086] Preferably, the live broadcast content is automatically analyzed and identified through the data evaluation module, and the data and content during the live broadcast are evaluated and scored. This can produce visual project comprehensive quality inspection results, as well as quality inspection result data management in multiple dimensions such as the live broadcast room, dealer stores, and regions, avoiding the need for staff to perform frequent statistical reviews when evaluating and viewing the live broadcast data.
[0087] Preferably, the data feedback module is used to collect and analyze evaluation feedback information of the overall live broadcast data, collect data on the accuracy and reliability of the content review system through user reports, complaints and other feedback methods, analyze user feedback information, and optimize the system's performance in identifying problems and misjudging problems.
[0088] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited to the above embodiments, and various changes can be made within the knowledge scope of those skilled in the art without departing from the spirit of the present invention.
Claims
1. A live broadcast platform content review intelligent evaluation system; characterized by: Included are: Data capture module: used to obtain video and audio data from live streams or uploaded media; Data preprocessing module: used to perform initial processing on the data captured by the data capture module; Data analysis module: used to extract and analyze the initially processed video and audio files; Data calculation module: used to perform quality inspection on the extracted data and determine whether it is compliant; Data storage module: used to back up and store the extracted and quality-inspected data; Data evaluation module: used to evaluate the overall live broadcast data indicators and live broadcast content; Data feedback module: used to allow users to report and provide feedback on content, and to conduct secondary confirmation of uncertain content.
2. According to claim 1, the content review intelligent evaluation system for a live broadcast platform is characterized in that: The data capture module is used to scan and extract data information of user live broadcast content or uploaded content. The data capture module can automatically obtain the content and related information of different live broadcast rooms in the live broadcast platform, and transmit the captured content and related information to the data preprocessing module.
3. According to claim 2, a live broadcast platform content review intelligent evaluation system is characterized by: The data capture module includes the following methods when capturing live broadcast information and content: A11: When capturing text data, directly capture the text data that can be directly collected from the bullet screen, comments, etc. on the live broadcast platform; A12: When capturing image data information, perform text extraction or intelligent recognition on live video streams, expressions, images and other related graphics and texts; A13: When extracting information from animations and interactive content, capture the information data in the video and audio by recording the screen.
4. According to claim 3, the content review intelligent evaluation system for a live broadcast platform is characterized in that: The data preprocessing module is used to perform preliminary processing on the live broadcast information and content captured by the data capture module, remove noise, inconsistent formats or other messy parts in the captured data, which can improve the accuracy and efficiency of subsequent data reading, while reducing the impact of interference parts on the sensitivity of the data analysis module, and transmit the preprocessed data to the data analysis module.
5. According to claim 4, a live broadcast platform content review intelligent evaluation system is characterized by: The data analysis module includes the following steps when analyzing the preprocessed data: S11: Obtain pre-processed live video information and content; S12: extracting the audio and text information in the video file; S13: Convert the recording file into a text file; S14: Analyze the keyword frequency and time.
6. According to claim 5, a live broadcast platform content review intelligent evaluation system is characterized by: The data calculation module includes the following steps when performing quality inspection on live content and information: S21: Create quality inspection rules, which include inspection scope, quality inspection regulations, time budget and time plan; S22: Perform quality inspection calculations to conduct multiple audit dimensions on the video stream, audio stream, bullet screen\comments, etc. in the live content; S23: After the quality inspection is completed, the quality inspection results and quality inspection report are obtained.
7. The live broadcast platform content review intelligent evaluation system according to claim 6, characterized in that: The data storage module includes an information resource management submodule and a cloud storage terminal. It integrates the captured information and quality inspection completed data, manages and stores the integrated data, and backs up the data storage to the cloud.
8. According to claim 7, a live broadcast platform content review intelligent evaluation system is characterized by: When the data storage module backs up and stores data, the following methods are included: A21: Create a master document based on the corresponding date on the cloud platform; A22: Create classification sub-documents in the main document according to the classification information; A23: Compress and encrypt the live broadcast data and content for backup; A24: The compressed and encrypted live broadcast data and content are placed in a sub-document.
9. The live broadcast platform content review intelligent evaluation system according to claim 8, characterized in that: The data evaluation module includes the following indicators when evaluating the overall live broadcast data indicators and content: A31: Data indicators, including the number of viewers in the live broadcast room, thousands of leads, live broadcast duration and other data indicators; A32: Keyword frequency, the number of times a keyword appears per hour in a live broadcast; A33: Number of keywords, the magnitude of the accumulated number of keywords in a live broadcast; A34: Oral speaking speed, counting the average number of oral speaking times per minute; A35: Silence. This refers to a period of long-term cessation of content output during live broadcast. A period of silence lasting up to three minutes is considered as one period of continuous silence.
10. A live broadcast platform content review intelligent evaluation system according to claim 9, characterized in that: The data feedback module is used to collect and analyze evaluation feedback information of the overall live broadcast data. Through user reports, complaints and other feedback methods, it collects data on the accuracy and reliability of the content review system, analyzes user feedback information, and optimizes the system's performance in identifying problems and misjudging problems.
