Multimedia content processing method, device, electronic device and storage medium

Abstract

The embodiments of the present application provide a multimedia content processing method, device, electronic device and storage medium, which relate to the fields of cloud technology and artificial intelligence technology. The method includes: obtaining each multimedia content that has passed the machine review and the respective traffic of each multimedia content; obtaining the residual rate of low-quality content after the machine review; determining the target manual review amount to be manually reviewed based on the target low-quality content residual rate and the target low-quality traffic residual rate and the low-quality content residual rate after the machine review; based on the respective traffic of each multimedia content, determining the target manual review amount of multimedia content in each multimedia content as the multimedia content to be manually reviewed. In the solution provided by the present application, the multimedia content to be manually reviewed changes with the change of traffic, does not require manual monitoring and observation, and can better meet the review needs. In addition, it can make the coverage of manual review wider and reduce the residual low-quality content.

Description

Technical Field

[0001] The present application relates to the fields of cloud technology and artificial intelligence technology. Specifically, the present application relates to a multimedia content processing method, device, electronic device and storage medium. Background Art

[0002] With the development of computer technology and mobile communication technology, various new media platforms have emerged. More and more users are publishing various multimedia data (also called multimedia content, such as short videos) through various new media platforms. In order to ensure the content quality of multimedia data, it is necessary to review the content of multimedia data and filter out multimedia data with unqualified content.

[0003] Currently, multimedia data is mostly reviewed using a semi-automated approach that combines machine review with manual review. This approach involves first conducting machine review, followed by manual review of the data that has passed the machine review. This approach requires a large amount of data to be manually reviewed after machine review. If all this data were to be manually reviewed, significant manpower and resources would be wasted. Furthermore, the sheer volume of manual review can result in suboptimal results, potentially leaving a significant amount of non-compliant multimedia data untouched. Summary of the Invention

[0004] The embodiments of the present application provide a multimedia content processing method, device, electronic device and storage medium. Based on this solution, the review efficiency and review effect of multimedia data can be effectively improved.

[0005] To achieve the above objectives, the specific technical solutions provided in the embodiments of the present application are as follows:

[0006] In one aspect, an embodiment of the present application provides a method for processing multimedia content, the method comprising:

[0007] Obtaining each multimedia content and its respective traffic volume that has been approved by the machine, wherein the traffic volume of each multimedia content is updated according to a preset traffic volume update cycle;

[0008] Obtain the residual rate of low-quality content after machine review, where the residual rate of low-quality content after machine review is the proportion of low-quality content in each historical multimedia content that has passed machine review;

[0009] The target amount of manual review to be performed is determined based on the target low-quality content residual rate, the target low-quality traffic residual rate, and the low-quality content residual rate after machine review. The target low-quality traffic residual rate is the preset traffic ratio of low-quality content in each multimedia content; the target low-quality content residual rate is the preset ratio of low-quality content in each multimedia content.

[0010] Based on the traffic of each multimedia content, a target manual review quantity of multimedia content is determined in each multimedia content as the multimedia content to be manually reviewed.

[0011] On the other hand, an embodiment of the present invention further provides a multimedia content processing device, the device comprising:

[0012] The first acquisition module is used to obtain each multimedia content that has been approved by the machine and the flow rate of each multimedia content, where the flow rate of each multimedia content is updated according to a preset time period;

[0013] The second acquisition module is used to obtain the low-quality content residual rate after machine review, where the low-quality content residual rate after machine review is the proportion of low-quality content in each historical multimedia content that has passed the machine review;

[0014] The first determination module is configured to determine a target amount of manual review to be performed based on a target low-quality content residual rate, a target low-quality traffic residual rate, and a low-quality content residual rate after machine review, wherein the target low-quality traffic residual rate is a preset traffic ratio of low-quality content in each multimedia content; and the target low-quality content residual rate is a preset ratio of low-quality content in each multimedia content;

[0015] The second determination module is configured to determine a target manual review quantity of multimedia content among the multimedia content based on the respective traffic of each multimedia content, as the multimedia content to be manually reviewed.

[0016] An embodiment of the present invention also provides an electronic device, which includes one or more processors; a memory; and one or more computer programs, wherein the one or more computer programs are stored in the memory and configured to be executed by the one or more processors, and the one or more computer programs are configured to execute the above-mentioned multimedia content processing method or the methods provided in various optional implementations of the multimedia content processing method.

[0017] An embodiment of the present invention further provides a computer-readable storage medium for storing a computer program. When the computer program runs on a processor, the processor can execute the above-mentioned multimedia content processing method or the methods provided in various optional implementations of the multimedia content processing method.

[0018] An embodiment of the present invention further provides a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the aforementioned multimedia content processing method or the methods provided in various optional implementations of the multimedia content processing method.

[0019] The beneficial effects of the technical solution provided by this application are:

[0020] The present application provides a multimedia content processing method, device, electronic device and storage medium. For each multimedia content that has passed the machine review, when determining the manual review amount, the target manual review amount to be performed is determined based on the target low-quality content residual rate and the target low-quality traffic residual rate and the low-quality content residual rate after the machine review. Since the target manual review amount takes into account the two dimensions of content and traffic, it can ensure to a certain extent that multimedia content with insufficient traffic can also be manually reviewed, thereby making the coverage of manual review wider and reducing the residual of low-quality content; in addition, the traffic of each multimedia content is updated according to the preset traffic cycle, so that the manually reviewed multimedia content changes dynamically with the change of traffic, without the need for manual monitoring and observation, and can better meet the review requirements. The technical solution of the present application improves the review effect of multimedia content, thereby improving the user experience when pushing multimedia content to users. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for describing the embodiments of the present application.

[0022] Figure 1 A flowchart of a multimedia content processing method provided in an embodiment of the present application;

[0023] Figure 2 A schematic diagram of the system architecture of the multimedia content review system provided in an embodiment of the present application;

[0024] Figure 3 A schematic diagram of the data processing process provided in an embodiment of the present application;

[0025] Figure 4 A schematic diagram of a process for determining the residual rate of low-quality content after machine review provided in an embodiment of the present application;

[0026] Figure 5 A schematic diagram of determining the target manual review amount provided in an embodiment of the present application;

[0027] Figure 6A schematic diagram of the structure of a multimedia content processing device provided in an embodiment of the present application;

[0028] Figure 7 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0029] The following describes in detail embodiments of the present application, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present application, and are not to be construed as limiting the present application.

[0030] It will be understood by those skilled in the art that, unless expressly stated otherwise, the singular forms "a", "an", "said" and "the" used herein may also include the plural forms. It should be further understood that the term "comprising" used in the specification of the present application refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof. It should be understood that when we refer to an element as being "connected" or "coupled" to another element, it may be directly connected or coupled to the other element, or there may be intermediate elements. In addition, "connected" or "coupled" as used herein may include wireless connections or wireless couplings. The term "and / or" used herein includes all or any units and all combinations of one or more associated listed items.

[0031] The solutions provided in the embodiments of the present application relate to fields such as cloud technology, big data, and artificial intelligence in computer technology. Specifically, the data processing, data calculation, data storage, etc. involved in the multimedia content processing method involved in the embodiments of the present application can be implemented through cloud technology (such as cloud computing, cloud storage, etc.). Cloud computing is a computing model that distributes computing tasks on a resource pool composed of a large number of computers, so that various application systems can obtain computing power, storage space, and information services as needed. The network that provides resources is called a "cloud". The resources in the "cloud" appear to be infinitely expandable to users, and can be obtained at any time, used on demand, expanded at any time, and paid for on a per-use basis.

[0032] As a provider of cloud computing infrastructure, a cloud computing resource pool (referred to as a cloud platform, generally referred to as an IaaS (Infrastructure as a Service) platform) is established. Various types of virtual resources are deployed in the resource pool for external customers to choose and use. The cloud computing resource pool mainly includes: computing devices (virtualized machines, including operating systems), storage devices, and network devices.

[0033] Based on logical functional divisions, the PaaS (Platform as a Service) layer can be deployed on top of the IaaS (Infrastructure as a Service) layer, and the SaaS (Software as a Service) layer can be deployed on top of the PaaS layer. SaaS can also be deployed directly on top of IaaS. PaaS is a platform for software execution, such as databases and web containers. SaaS is a variety of business software, such as web portals and text messaging tools. Generally speaking, SaaS and PaaS are upper layers relative to IaaS.

[0034] Cloud computing refers to the delivery and usage model of IT infrastructure, enabling on-demand, scalable access to required resources over the internet. Broadly speaking, cloud computing refers to the delivery and usage model of services, enabling on-demand, scalable access to required services over the internet. These services can be IT-related, software-related, internet-related, or other services. Cloud computing is the product of the convergence of traditional computer and network technologies, including grid computing, distributed computing, parallel computing, utility computing, network storage technologies, virtualization, and load balancing.

[0035] Cloud computing has rapidly grown, driven by the internet, real-time data streams, the diversification of connected devices, and the growing demand for search services, social networks, mobile commerce, and open collaboration. Unlike previous parallel and distributed computing approaches, the emergence of cloud computing will fundamentally revolutionize the entire internet and enterprise management model.

[0036] The training of the machine audit model involved in the embodiments of the present application can be achieved through machine learning in artificial intelligence technology.

[0037] Artificial Intelligence (AI) refers to the theories, methods, techniques, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, to perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results. In other words, AI is a comprehensive technology within computer science that seeks to understand the essence of intelligence and produce new intelligent machines that can respond in a manner similar to human intelligence. AI also involves studying the design principles and implementation methods of various intelligent machines, enabling them to possess the capabilities of perception, reasoning, and decision-making.

[0038] Artificial intelligence technology is a comprehensive discipline covering a wide range of fields, encompassing both hardware and software technologies. Basic AI technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data and multimedia content processing, operating / interactive systems, and mechatronics. The AI ​​technologies involved in the embodiments of this application primarily encompass natural language processing, machine learning, and deep learning.

[0039] Natural language processing (NLP) is a key area of ​​research in computer science and artificial intelligence. It studies the theories and methods that enable effective communication between humans and computers using natural language. Natural language processing (NLP) integrates linguistics, computer science, and mathematics. Therefore, research in this field involves natural language—the language we use in everyday life—and is closely linked to the study of linguistics. Natural language processing technologies typically include text processing, semantic understanding, machine translation, robotic question answering, and knowledge graphs.

[0040] Machine learning (ML) is a multidisciplinary field that encompasses probability theory, statistics, approximation theory, convex analysis, and algorithmic complexity theory. It specifically studies how computers can simulate or implement human learning behaviors to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance. Machine learning is at the core of artificial intelligence and the fundamental way to make computers intelligent. Its applications span all areas of AI. Machine learning and deep learning typically include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and self-learning.

[0041] The training data required for the model training of the machine audit model involved in the embodiments of the present application can be big data obtained from the Internet.

[0042] Big data refers to collections of data that cannot be captured, managed, and processed within a specific timeframe using conventional software tools. These massive, rapidly growing, and diverse information assets require new processing models to enhance decision-making, insight discovery, and process optimization. With the advent of the cloud era, big data has attracted increasing attention. Big data requires specialized technologies to efficiently process large amounts of time-sensitive data. Technologies suitable for big data include massively parallel processing databases, data mining, distributed file systems, distributed databases, cloud computing platforms, the internet, and scalable storage systems.

[0043] The following specific embodiments describe in detail the technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0044] The executors of the technical solution of this application are computer devices, including but not limited to servers, personal computers, laptops, tablet computers, smart phones, etc. Computer devices include user devices and network devices. User devices include but are not limited to computers, smart phones, PADs, etc. Network devices include but are not limited to a single network server, a server group consisting of multiple network servers, or a cloud composed of a large number of computers or network servers in cloud computing, wherein cloud computing is a type of distributed computing, a super virtual computer composed of a group of loosely coupled computer sets. Computer devices can be run alone to implement this application, or they can be connected to the network and implement this application through interactive operations with other computer devices in the network. The network where the computer device is located includes but is not limited to the Internet, wide area network, metropolitan area network, local area network, VPN network, etc.

[0045] The embodiment of the present application provides a multimedia content processing method, the execution subject of the method can be any electronic device, for example, the method can be executed by a server, such as Figure 1 As shown, the method may include:

[0046] Step S101, obtaining each multimedia content that has passed machine review and the traffic of each multimedia content;

[0047] Multimedia content includes, but is not limited to, text, images, a combination of images and text, short videos, long videos, etc. Multimedia content can include user-generated content (UGC), professionally produced content (PGC), multi-channel network (MCN), and professional user-generated content (PUGC).

[0048] The multimedia content may be any multimedia content that needs to be reviewed. In some optional embodiments, the multimedia content may be multimedia content sent by a user through a user terminal and received by a server corresponding to the multimedia publishing platform, or it may be multimedia content obtained by a server corresponding to the multimedia publishing platform from a preset storage space. Due to the large amount of multimedia content to be reviewed, in order to improve the review efficiency and ensure the review quality, the review is conducted by combining machine review with manual review. After the multimedia content is obtained, it is first reviewed by a machine. For the multimedia content that has passed the machine review, a portion of it is selected for manual review, thereby reducing the workload of manual review.

[0049] Machine review involves a computer utilizing a machine review model to review multimedia content based on the characteristics of each dimension of the multimedia content, reviewing whether each characteristic of the multimedia content meets pre-set quality criteria, and determining whether the multimedia content is qualified based on the evaluation results of each characteristic. The characteristics of the multimedia content may include characteristics that characterize the quality of the multimedia content, such as image clarity, as well as characteristics such as timeliness and tone (positive, negative, or neutral).

[0050] Among them, the specific model structure of the machine audit model is not limited in the embodiments of this application. In some optional embodiments, the audit model can be a multimodal machine learning (MMML) model, for example, a convolutional neural network model. The various features of the multimedia content can be extracted by the neural network model, and the multimedia content can be classified based on the various features. The audit results of the multimedia content can be determined according to the classification results (such as qualified or unqualified quality). Therefore, based on the results of the machine audit, the multimedia content that has passed the machine audit can be determined, so that the multimedia content that has been manually audited can be further screened out from the multimedia content that has passed the machine audit through subsequent processing methods.

[0051] When the server obtains each piece of multimedia content approved by the machine, it can obtain the traffic volume of each piece of multimedia content at each preset time point. By calculating the traffic volume of each piece of multimedia data at each preset time point, it can obtain the total traffic volume of each piece of multimedia content approved by the machine within the preset time range. In one example, the multimedia content is a short video. Based on the number of views of each short video in the past hour, the server can calculate the total number of views of each short video in the past 24 hours.

[0052] In addition, a traffic update period is pre-configured, and the traffic of each multimedia content is updated according to the pre-configured traffic update period to ensure the timeliness of traffic changes, which serves as the basis for subsequent determination of multimedia content for manual review. In one example, the total traffic of each multimedia content in the past 24 hours is obtained every 15 minutes.

[0053] Step S102, obtaining the residual rate of low-quality content after machine review;

[0054] Among them, the residual rate of low-quality content after machine review is the proportion of low-quality content in each historical multimedia content that has passed the machine review to the total low-quality content, that is, the inaccuracy rate of the machine review. Among them, historical multimedia content is the historical data before processing the current multimedia content, and the residual rate of low-quality content after machine review is determined based on the historical data. Low-quality content is multimedia content that does not meet the preset quality conditions, that is, content with unqualified quality. Low quality, low quality

[0055] In one possible implementation, obtaining the residual rate of low-quality content after machine review includes:

[0056] Obtaining each historical multimedia content that has passed machine review, and performing manual review on each historical multimedia content to obtain historical manual review results;

[0057] Based on historical manual review results, determine the residual rate of low-quality content after machine review.

[0058] In actual applications, the audit model for machine review can be determined in advance, and a batch of test samples (that is, historical multimedia content) can be input into the audit model. Each test sample can include non-low-quality multimedia content and low-quality multimedia content. The multimedia content output by the audit model is then manually reviewed. Based on the proportion of low-quality multimedia content manually reviewed, the proportion of low-quality content in each historical multimedia content that has passed the machine review is determined, and the low-quality content residual rate after review by the audit model is obtained.

[0059] For example, 200 multimedia contents are machine-reviewed, and 60 low-quality multimedia contents are found by the machine. The 140 multimedia contents that passed the machine review are then manually reviewed, and the number of low-quality multimedia contents found by manual review is 40. The interception rate of low-quality contents manually reviewed is 40 ÷ (40 + 60) = 40%, and the residual rate of low-quality contents after machine review is 40%.

[0060] In a possible implementation, the method further includes:

[0061] Based on the preset residual rate update cycle, the residual rate of low-quality content after machine review is updated.

[0062] In actual applications, the residual rate of low-quality content after machine review can be updated regularly based on a preset residual rate update cycle. According to the updated residual rate of low-quality content, the amount of multimedia content to be manually reviewed in the multimedia content currently to be processed is determined, which can ensure that the multimedia content to be manually reviewed is continuously updated, thereby ensuring the timeliness of the review.

[0063] Step S103 , determining a target manual review amount to be performed based on the target low-quality content residual rate, the target low-quality traffic residual rate, and the low-quality content residual rate after machine review.

[0064] Among them, the target low-quality content residual rate is the proportion of low-quality content in each preset multimedia content, that is, after machine review and manual review, the ratio of the number of residual low-quality multimedia content to the number of all low-quality multimedia content. The target low-quality traffic residual rate is the traffic ratio of low-quality content in each preset multimedia content, which means that after machine review and manual review, the ratio of the traffic of residual low-quality multimedia content to the traffic of all low-quality multimedia content. Among them, traffic can refer to the number of times the multimedia content has been viewed by the user. For example, traffic can be the amount of playback or reading of multimedia content, etc., which is not limited in this embodiment. The target low-quality content residual rate and the target low-quality traffic residual rate can both be determined based on actual needs, experience values ​​and / or experimental values. For example, the target low-quality content residual rate can be 5%, and the target low-quality traffic residual rate can be 1%.

[0065] The target manual review volume is the amount of multimedia content that requires manual review. This target manual review volume can be less than the total amount of multimedia content that has passed machine review. The remaining low-quality content rate after machine review, the target low-quality content rate, and the target low-quality traffic rate are used to calculate the amount of multimedia content that still requires manual review after passing machine review.

[0066] In one possible implementation, a target amount of manual review to be performed is determined based on the target low-quality content residual rate, the target low-quality traffic residual rate, and the low-quality content residual rate after machine review, including:

[0067] Determining a first amount of multimedia content to be manually reviewed based on a low-quality content residual rate after machine review and a target low-quality content residual rate;

[0068] Determining a second amount of multimedia content to be manually reviewed based on the low-quality content residual rate after machine review and the target low-quality traffic residual rate;

[0069] A target manual review amount is determined based on a larger value of the first quantity and the second quantity.

[0070] In practical applications, the amount of multimedia content to be manually reviewed in the content dimension, i.e., the first amount, can be determined based on the low-quality content residual rate after machine review and the target low-quality content residual rate. The amount of multimedia content to be manually reviewed in the traffic dimension, i.e., the second amount, can also be determined based on the low-quality content residual rate after machine review and the target low-quality traffic residual rate. To ensure the coverage of the review, the amount of multimedia content to be manually reviewed is determined based on the larger of the first and second amounts. Optionally, the target manual review amount can be greater than or equal to the larger of the first and second amounts.

[0071] In the embodiment of the present application, the amount of multimedia content to be manually reviewed is determined based on the two dimensions of content and traffic. This can ensure that multimedia content with insufficient traffic can also be manually reviewed, thereby expanding the coverage of manual review, reducing the residual of low-quality content, and improving the user experience.

[0072] In one possible implementation, determining a first amount of multimedia content to be manually reviewed based on a low-quality content residual rate after machine review and a target low-quality content residual rate includes:

[0073] Determine the proportion of multimedia content to be manually reviewed based on the low-quality content residual rate after machine review and the target low-quality content residual rate;

[0074] Obtaining the quantity of each multimedia content;

[0075] Based on the quantity ratio and the quantity of each multimedia content, a first quantity of multimedia content to be manually reviewed is determined.

[0076] In actual applications, the proportional relationship between the number of multimedia contents to be manually reviewed and the number of multimedia contents passed by the machine review can be determined based on the residual rate of low-quality content after machine review and the target residual rate of low-quality content. According to this proportional relationship and the number of multimedia contents passed by the machine review, the number of multimedia contents to be manually reviewed in the content dimension can be determined, that is, the first number.

[0077] In one example, the quantity ratio is determined according to the following formula (1):

[0078]

[0079] Where X represents the percentage of multimedia content subject to manual review among all multimedia content after machine review. A represents the target low-quality content retention rate. Q represents the low-quality content retention rate after machine review.

[0080] In one possible implementation, determining the second amount of multimedia content to be manually reviewed based on the low-quality content residual rate after machine review and the target low-quality traffic residual rate includes:

[0081] Based on the low-quality content residual rate after machine review and the target low-quality traffic residual rate, determine the traffic proportion of multimedia content to be manually reviewed in each multimedia content;

[0082] Obtaining traffic of various multimedia contents;

[0083] The second quantity of multimedia contents to be manually reviewed is determined based on the traffic ratio of the multimedia contents to be manually reviewed in the multimedia contents and the traffic of each multimedia content.

[0084] In practical applications, the ratio of the total traffic volume of multimedia content subject to manual review to the total traffic volume of each multimedia content subject to machine review can be determined based on the residual rate of low-quality content after machine review and the target residual rate of low-quality traffic volume. The traffic volume of each multimedia content is obtained, and the total traffic volume of each multimedia content is calculated. Based on the traffic volume ratio and the total traffic volume of each multimedia content subject to machine review, the number of multimedia contents subject to manual review (i.e., the second number) can be determined in terms of traffic volume.

[0085] In one example, the traffic proportion is determined according to the following formula (2):

[0086]

[0087] Where Y represents the traffic ratio of multimedia content to be manually reviewed in the total volume of multimedia content after machine review. B represents the target low-quality traffic residual rate. Q represents the low-quality content residual rate after machine review.

[0088] In one example, the number of multimedia contents that have been reviewed by the machine is 20. Figure 5 As shown in the figure, content 1, content 2...content 20 are sorted by the flow rate of each multimedia content. According to the residual rate of low-quality content after machine review and the target residual rate of low-quality content, the proportion X of the number of multimedia contents to be manually reviewed in each multimedia content is determined to be 50%. According to the number of multimedia contents 20, the first number of multimedia contents to be manually reviewed is obtained as 20×50%=10. According to the residual rate of low-quality content after machine review and the target residual rate of low-quality content, the flow rate proportion Y of multimedia contents to be manually reviewed in each multimedia content is determined to be 25%. According to the total flow rate of multimedia content 1000, the flow rate of multimedia contents to be manually reviewed is obtained as 250. The total flow rate of the first 5 multimedia contents in the 20 multimedia contents is 250, and the second number is 5. According to the larger value of the first number and the second number, the target manual review amount of the multimedia content is determined. The target manual review amount can be 10. The flow rate of the 10th multimedia content in the sorted list, that is, content 10 (as shown in the video playback number (video view, VV) value) is used as the traffic threshold. When manual review is performed, multimedia content with a traffic greater than or equal to the VV value will be manually reviewed, and multimedia content that passes the manual review will be pushed to the user. Multimedia content with a traffic less than the VV value will not be manually reviewed and the content will be pushed directly.

[0089] Step S104 : Based on the flow rates of the multimedia contents, a target number of multimedia contents for manual review is determined from the multimedia contents as the multimedia contents to be manually reviewed.

[0090] After the server determines the target manual review volume and the traffic volume of each multimedia content that has passed the machine review within a preset time range, it determines the target manual review volume of multimedia content from all multimedia content that has passed the machine review based on the traffic volume of each multimedia content, and performs manual review on these multimedia content. For example, these multimedia content can be sent to a manual review pool so that relevant personnel can manually review these multimedia content. Since the traffic volume of each multimedia content is updated according to a preset time period and changes dynamically, the multimedia content for manual review determined based on the traffic volume changes dynamically with the change in traffic volume, and there is no need for manual monitoring and observation to determine the multimedia content for manual review.

[0091] The multimedia content that has passed the manual review can be stored in the push database for subsequent content push. For multimedia content that has not passed the manual review, it means that these multimedia contents are of low quality and are not suitable for push.

[0092] The multimedia content processing method provided in the embodiment of the present application, for each multimedia content that has passed the machine review, when determining the manual review amount, determines the target manual review amount to be performed based on the target low-quality content residual rate and the target low-quality traffic residual rate and the low-quality content residual rate after machine review. Since the target manual review amount takes into account the two dimensions of content and traffic, it can ensure to a certain extent that multimedia content with insufficient traffic can also be manually reviewed, thereby making the coverage of manual review wider and reducing the residual of low-quality content; in addition, the traffic of each multimedia content is updated according to the preset traffic cycle, so that the multimedia content that is manually reviewed changes dynamically with the change of traffic, does not require manual monitoring and observation, and can better meet the review requirements. The technical solution of the present application improves the review effect of multimedia content, thereby improving the user experience when pushing multimedia content to users.

[0093] In a possible implementation, based on the traffic of each multimedia content, determining a target manual review quantity of multimedia content from each multimedia content as the multimedia content to be manually reviewed includes:

[0094] A target number of multimedia contents for manual review are selected in descending order of the flow rates of the multimedia contents as the multimedia contents to be manually reviewed.

[0095] In actual applications, after the server obtains the traffic of each multimedia content that has passed the machine review, it selects the target number of multimedia contents for manual review from the multimedia contents that have passed the machine review in order of traffic from large to small for manual review.

[0096] Optionally, the multimedia contents are sorted by traffic volume, and a traffic threshold is determined based on the amount of manual review and the traffic volume of the sorted multimedia contents. The traffic threshold may be the traffic volume of one of the multimedia contents. When determining the multimedia contents for manual review, multimedia contents with traffic volume greater than or equal to the traffic threshold may be selected as the multimedia contents to be manually reviewed.

[0097] Since the traffic of multimedia content is updated according to a preset period, the traffic threshold changes dynamically with the update of the traffic, and does not require manual monitoring and observation. As a result, the multimedia content that is manually reviewed based on the traffic threshold changes dynamically, which can better meet the review requirements and improve the review effect.

[0098] Since multimedia content with larger traffic volume will be pushed first in subsequent content push, selecting manually reviewed multimedia data in order of traffic volume from large to small can ensure to a certain extent that the pushed content has been manually reviewed, thereby ensuring that the pushed content meets the requirements in terms of quality, timeliness, etc., thereby improving user experience.

[0099] In a possible implementation, selecting multimedia contents with a target manual review volume in descending order of the respective traffic volumes of the multimedia contents as the multimedia contents to be manually reviewed includes any of the following:

[0100] Determine multimedia content that has not been manually reviewed from the target number of multimedia content for manual review, and use the multimedia content that has not been manually reviewed as multimedia content to be manually reviewed; or

[0101] Multimedia contents that have not been manually reviewed are determined from the multimedia contents, and a target number of multimedia contents that have not been manually reviewed are selected from the multimedia contents that have not been manually reviewed as multimedia contents to be manually reviewed.

[0102] In actual applications, after determining the target amount of multimedia content for manual review among the multimedia content that has passed the machine review, these multimedia contents may have been manually reviewed before. In order to avoid repeated review, these multimedia contents that have been manually reviewed can be filtered out to obtain multimedia content that has not been manually reviewed, and the multimedia content that has not been manually reviewed can be reviewed, or, the multimedia content that has been manually reviewed can be filtered out from the multimedia content that has passed the machine review, and a part of the multimedia content that has not been manually reviewed can be selected for manual review.

[0103] The following describes the system architecture of the technical solution of the present application in detail through a specific embodiment. This embodiment is only one embodiment of the technical solution of the present application and does not represent all implementation methods of the technical solution of the present application.

[0104] like Figure 2 As shown, this embodiment provides a multimedia content review system, which includes user terminals 11, 12, ..., 1n, a server 2, and a manual review terminal 3. In this embodiment, the multimedia content is a short video, and the traffic of the multimedia content is the playback volume of the short video.

[0105] User terminals 11, 12, ..., 1n each obtain short videos. These short videos can be short videos shot by the user through the user terminal or short videos downloaded to the user terminal. These short videos include short videos with qualified content and short videos with unqualified content (i.e., low quality), totaling 100 short videos. User terminals 11, 12, ..., 1n send these short videos to server 2 corresponding to the short video publishing platform. Server 2 verifies the content of each short video using an audit model. Specifically, the audit model obtains the quality characteristics (e.g., clarity, duration) and timeliness characteristics of each short video. Based on the quality characteristics and timeliness characteristics of each short video, each short video is audited. After the audit model audits, 20% of the short videos are low-quality. The audit model intercepts these short videos with low-quality content. 80% of the short videos with qualified content are short videos that have passed the machine audit. The number of short videos that have passed the machine audit is 100 × 80% = 80. The audit qualification rate of the audit model determined in advance based on each test sample of the audit model is 80%, then the residual rate of low-quality content after machine audit is 1-60%=20%, the target low-quality content residual rate of low-quality content of short videos pre-configured for human-machine audit is 5%, and the playback volume residual rate corresponding to the target low-quality content is 1%. Based on the low-quality content residual rate after machine audit and the target low-quality content residual rate, the proportion of short videos to be manually audited is calculated: 1-5% / 20%=0.75, so the first number of short videos to be manually audited is 80×0.75=60; based on the low-quality content residual rate after machine audit and the target low-quality traffic residual rate, the traffic proportion of short videos to be manually audited is calculated: 1-1% / 20%=0.95, the total playback volume of 80 short videos passed by machine audit is 1000, then the total playback volume of multimedia data that requires manual audit is 1000×0.95=950, and these 80 short videos can be sorted from large to small according to the playback volume. , select the number of short videos corresponding to the playback volume of 950, and add the playback volume of the short videos ranked at the top in sequence, and take the short video corresponding to the first time the sum of the playback volume is equal to or greater than 950, and the sum of the number of videos of the short videos ranked before the short video as the second number. Assuming that the total playback volume of the first 50 short videos is 950, accounting for 95% of the 80 short videos, the second number is determined to be 50, and the larger number between 50 and 60, that is, 60, is taken, and it is determined that the number of short videos to be manually reviewed is greater than or equal to 60. After sorting by playback volume, the short videos ranked in the top 61 of the 80 short videos are selected, and the playback volume of the short video ranked 61 is 10. When manually reviewing the short videos, the manual review terminal 3 reviews the short videos with a playback volume greater than or equal to 10, and sends the short videos that pass the review to the server 2 for content push. Other short videos with smaller traffic are not manually reviewed, and the server directly pushes content.

[0106] The data processing process in the technical solution of this application is described in detail below through a specific embodiment.

[0107] like Figure 3As shown, a user terminal obtains multimedia content and sends it to the server corresponding to the multimedia content publishing platform (shown as "Post" in the figure). The server first performs a security review of the multimedia content based on preset security review rules. Security review rules can be configured according to specific needs. For example, multimedia content involving pornography does not comply with security review rules and will be blocked during the security review. It is then reviewed to see if it complies with basic rules. Basic rules can be the rules specified by the multimedia publishing platform for publishing multimedia content. For example, if the multimedia content is a short video and the multimedia publishing platform stipulates a playback time of 20 seconds for short videos, short videos with a playback time exceeding 20 seconds do not comply with the basic rules. During the basic rule review, short videos with a playback time exceeding 20 seconds will be blocked. Short videos that have passed the security review and basic rule review enter the human-machine review link. The machine (shown as the "Feature Engine") adds features to each multimedia content. To facilitate subsequent content push, the features added by the machine to each multimedia content can be content tags. For example, if the multimedia content is a video of a TV series, the tags added by the machine to the multimedia content can be the title of the TV series, the name of the actor, etc. The machine extracts the quality features of multimedia content, uses the audit model to perform a quality audit on the multimedia content (as shown in the "Quality Judgment" in the figure), and determines whether the multimedia content is repeated with the multimedia content in the preset database (as shown in the "Duplicate Elimination" in the figure). If repeated, the identifier of the stored duplicate multimedia content is used as the identifier of the duplicate multimedia content to avoid repeated processing of the multimedia content. After passing the machine review, some low-quality multimedia content is intercepted. For the multimedia content that has passed the machine review, the low-quality content residual rate after the review by the review model is determined in advance based on the test samples of the review model. According to the low-quality content residual rate after the review by the review model and the pre-configured target low-quality content residual rate, the number of multimedia content to be manually reviewed in the content dimension is calculated. According to the low-quality content residual rate after the review by the review model and the pre-configured target low-quality traffic residual rate, the number of multimedia content to be manually reviewed in the traffic dimension is calculated. The target manual review volume of the multimedia content is determined based on the larger number of multimedia content to be manually reviewed corresponding to the content dimension and the traffic dimension. The traffic of each multimedia content is obtained. According to the order of traffic from large to small, the target manual review volume of multimedia content is determined among the multimedia content that has passed the machine review. This part of the multimedia content is sent for manual review through a trigger (as shown in the "human review" in the figure). The review results of the manual review are fed back to the machine review model (as shown in the "online feedback" in the figure), and the low-quality content residual rate of the machine review model is updated.

[0108] The following is a detailed description of the process of determining the residual rate of low-quality content reviewed by the machine in the technical solution of this application through a specific embodiment.

[0109] like Figure 4 As shown, the user terminal obtains multimedia content and sends the multimedia content to the server corresponding to the multimedia content publishing platform (as shown in the "Publish" in the figure). The server uses the audit model to audit the multimedia content (as shown in the "Machine Audit" in the figure). Specifically, the audit model obtains the quality characteristics (for example, clarity, time length) and timeliness characteristics of each multimedia content. Based on the quality characteristics and timeliness characteristics of each multimedia content, each multimedia content is audited. After auditing by the audit model, low-quality multimedia content is intercepted, and qualified multimedia content passes the audit and enters manual audit (as shown in the "Human Audit" in the figure). Assuming that the human audit can audit all low-quality multimedia content, the proportion of low-quality multimedia content audited by humans in the total amount of low-quality multimedia content, that is, the proportion of low-quality multimedia content not audited by the machine audit in the total amount of low-quality multimedia content, is the low-quality content residual rate after machine audit. Multimedia content that passes manual audit can be recommended (as shown in the "Enter Recommendation" in the figure). Multimedia content that fails manual audit is intercepted and the process terminates.

[0110] The multimedia content processing method provided in the embodiment of the present application, for each multimedia content that has passed the machine review, when determining the manual review amount, determines the target manual review amount to be performed based on the target low-quality content residual rate and the target low-quality traffic residual rate and the low-quality content residual rate after machine review. Since the target manual review amount takes into account the two dimensions of content and traffic, it can ensure to a certain extent that multimedia content with insufficient traffic can also be manually reviewed, thereby making the coverage of manual review wider and reducing the residual of low-quality content; in addition, the traffic of each multimedia content is updated according to the preset traffic cycle, so that the multimedia content that is manually reviewed changes dynamically with the change of traffic, does not require manual monitoring and observation, and can better meet the review requirements. The technical solution of the present application improves the review effect of multimedia content, thereby improving the user experience when pushing multimedia content to users.

[0111] and Figure 1 Based on the same principle as the method shown in , the embodiment of the present disclosure further provides a multimedia content processing device 20, such as Figure 6 As shown, the multimedia content processing device 20 includes:

[0112] The first acquisition module 21 is used to obtain each multimedia content that has been approved by the machine and the flow rate of each multimedia content. The flow rate of each multimedia content is updated according to a preset time period;

[0113] The second acquisition module 22 is used to obtain the low-quality content residual rate after machine review, where the low-quality content residual rate after machine review is the proportion of low-quality content in each historical multimedia content that has passed the machine review;

[0114] The first determination module 23 is configured to determine a target amount of manual review to be performed based on a target low-quality content residual rate, a target low-quality traffic residual rate, and a low-quality content residual rate after machine review, wherein the target low-quality traffic residual rate is a preset traffic ratio of low-quality content in each multimedia content; and the target low-quality content residual rate is a preset ratio of low-quality content in each multimedia content;

[0115] The second determining module 24 is configured to determine a target manual review quantity of multimedia contents from among the multimedia contents based on the respective traffic of the multimedia contents, as the multimedia contents to be manually reviewed.

[0116] In a possible implementation, the first determining module 23 is configured to:

[0117] Determining a first amount of multimedia content to be manually reviewed based on a low-quality content residual rate after machine review and a target low-quality content residual rate;

[0118] Determining a second amount of multimedia content to be manually reviewed based on the low-quality content residual rate after machine review and the target low-quality traffic residual rate;

[0119] A target manual review amount is determined based on a larger value of the first quantity and the second quantity.

[0120] In one possible implementation, when determining the first amount of multimedia content to be manually reviewed based on the low-quality content residual rate after machine review and the target low-quality content residual rate, the first determination module 23 is configured to:

[0121] Determine the proportion of multimedia content to be manually reviewed based on the low-quality content residual rate after machine review and the target low-quality content residual rate;

[0122] Obtaining the quantity of each multimedia content;

[0123] Based on the quantity ratio and the quantity of each multimedia content, a first quantity of multimedia content to be manually reviewed is determined.

[0124] In one possible implementation, when determining the second amount of multimedia content to be manually reviewed based on the low-quality content residual rate after machine review and the target low-quality traffic residual rate, the first determination module 23 is configured to:

[0125] Based on the low-quality content residual rate after machine review and the target low-quality traffic residual rate, determine the traffic proportion of multimedia content to be manually reviewed in each multimedia content;

[0126] Obtaining traffic of various multimedia contents;

[0127] The second quantity of multimedia contents to be manually reviewed is determined based on the traffic ratio of the multimedia contents to be manually reviewed in the multimedia contents and the traffic of each multimedia content.

[0128] In a possible implementation, the second determining module 24 is configured to:

[0129] A target number of multimedia contents for manual review are selected in descending order of the flow rates of the multimedia contents as the multimedia contents to be manually reviewed.

[0130] In a possible implementation, the second determining module 24 selects the target manual review amount multimedia contents as the multimedia contents to be manually reviewed in descending order of the traffic of each multimedia content, and performs any of the following:

[0131] Determine multimedia content that has not been manually reviewed from the target number of multimedia content for manual review, and use the multimedia content that has not been manually reviewed as multimedia content to be manually reviewed; or

[0132] Multimedia contents that have not been manually reviewed are determined from the multimedia contents, and a target number of multimedia contents for manual review are selected from the multimedia contents that have not been manually reviewed as multimedia contents to be manually reviewed.

[0133] In a possible implementation, the first obtaining module 21 is specifically configured to:

[0134] Obtaining each historical multimedia content that has passed machine review, and performing manual review on each historical multimedia content to obtain historical manual review results;

[0135] Based on historical manual review results, determine the residual rate of low-quality content after machine review.

[0136] In a possible implementation, the multimedia content processing device 20 further includes:

[0137] The update module is used to update the residual rate of low-quality content after machine review based on a preset residual rate update cycle.

[0138] The multimedia content processing device of the embodiment of the present disclosure can execute the multimedia content processing device provided by the embodiment of the present disclosure. Figure 1The corresponding multimedia content processing method has a similar implementation principle. The actions performed by each module in the multimedia content processing device in the embodiment of the present disclosure correspond to the steps in the multimedia content processing method in the embodiment of the present disclosure. For a detailed functional description of each module in the multimedia content processing device, please refer to the description of the corresponding multimedia content processing method shown in the previous text, and will not be repeated here.

[0139] The multimedia content processing device provided in the embodiment of the present application determines the target manual review amount for each multimedia content that has passed the machine review, based on the target low-quality content residual rate and the target low-quality traffic residual rate and the low-quality content residual rate after machine review. Since the target manual review amount takes into account the two dimensions of content and traffic, it can ensure to a certain extent that multimedia content with insufficient traffic can also be manually reviewed, thereby making the coverage of manual review wider and reducing the residual of low-quality content; in addition, the traffic of each multimedia content is updated according to the preset traffic cycle, so that the multimedia content that is manually reviewed changes dynamically with the change of traffic, does not require manual monitoring and observation, and can better meet the review requirements. The technical solution of the present application improves the review effect of multimedia content, thereby improving the user experience when pushing multimedia content to users.

[0140] Among them, the multimedia content processing device can be a computer program (including program code) running in a computer device, for example, the multimedia content processing device is an application software; the device can be used to execute the corresponding steps in the method provided in the embodiment of the present application.

[0141] In some embodiments, the multimedia content processing device provided by the embodiments of the present invention can be implemented using a combination of software and hardware. As an example, the multimedia content processing device provided by the embodiments of the present invention can be a processor in the form of a hardware decoding processor, which is programmed to execute the multimedia content processing method provided by the embodiments of the present invention. For example, the processor in the form of a hardware decoding processor can adopt one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field-programmable gate arrays (FPGAs), or other electronic components.

[0142] In other embodiments, the multimedia content processing device provided by the embodiment of the present invention can be implemented in software. Figure 6 A multimedia content processing device stored in a memory is shown, which can be software in the form of a program and a plug-in, and includes a series of modules, including a first acquisition module 21, a second acquisition module 22, a first determination module 23, and a second determination module 24, for implementing the multimedia content processing method provided in an embodiment of the present invention.

[0143] The above embodiment introduces a multimedia content processing device from the perspective of a virtual module. The following describes an electronic device from the perspective of a physical module, as shown below:

[0144] The present application embodiment provides an electronic device, such as Figure 7 As shown, Figure 7 The electronic device 7000 shown includes: a processor 7001 and a memory 7003. The processor 7001 and the memory 7003 are connected, for example, via a bus 7002. Optionally, the electronic device 7000 may further include a transceiver 7004. It should be noted that in actual applications, the number of transceivers 7004 is not limited to one, and the structure of the electronic device 7000 does not constitute a limitation on the embodiments of the present application.

[0145] Processor 7001 may be a CPU, a general-purpose processor, a GPU, a DSP, an ASIC, an FPGA, or other programmable logic device, a transistor logic device, a hardware component, or any combination thereof. It may implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. Processor 7001 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, and the like.

[0146] The bus 7002 may include a path for transmitting information between the above components. The bus 7002 may be a PCI bus or an EISA bus, etc. The bus 7002 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 7 Only one thick line is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0147] The memory 7003 may be a ROM or other type of static storage device that can store static information and instructions, a RAM or other type of dynamic storage device that can store information and instructions, or an EEPROM, a CD-ROM or other optical disk storage, an optical disc storage (including a compact disc, a laser disc, an optical disc, a digital versatile disc, a Blu-ray disc, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited to these.

[0148] The memory 7003 is used to store application code for executing the solution of the present application, and the execution is controlled by the processor 7001. The processor 7001 is used to execute the application code stored in the memory 7003 to implement the content shown in any of the above method embodiments.

[0149] An embodiment of the present application provides an electronic device, and the electronic device in the embodiment of the present application includes: one or more processors; a memory; one or more computer programs, wherein the one or more computer programs are stored in the memory and configured to be executed by the one or more processors, and when the one or more programs are executed by the processor, each multimedia content that has passed the machine review and the respective traffic of each multimedia content are obtained, and the respective traffic of each multimedia content is updated according to a preset traffic update period; the low-quality content residual rate after machine review is obtained, and the low-quality content residual rate after machine review is the proportion of low-quality content in each historical multimedia content that has passed the machine review; according to the target low-quality content residual rate and the target low-quality traffic residual rate and the low-quality content residual rate after machine review, the target manual review amount to be manually reviewed is determined, the target low-quality traffic residual rate is the preset traffic ratio of low-quality content in each multimedia content; the target low-quality content residual rate is the preset ratio of low-quality content in each multimedia content; based on the respective traffic of each multimedia content, a target manual review amount of multimedia content is determined in each multimedia content as the multimedia content to be manually reviewed.

[0150] An embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon. When the computer program runs on a processor, the processor can execute the corresponding contents of the aforementioned method embodiment.

[0151] According to one aspect of the present application, a computer program product or computer program is provided, comprising computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the methods provided in various optional implementations of the aforementioned multimedia content processing method.

[0152] It should be understood that although the steps in the flowcharts of the accompanying drawings are shown in sequence as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some of the steps in the flowcharts of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.

[0153] The above descriptions are only partial embodiments of the present invention. It should be pointed out that ordinary technicians in this technical field can make several improvements and modifications without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A multimedia content processing method, applied to multimedia content review, wherein the multimedia content review includes machine review and manual review, characterized in that: The method comprises: Acquire each multimedia content that has been approved by the machine and the flow rate of each multimedia content, wherein the flow rate of each multimedia content is updated according to a preset flow rate update period; Obtaining a low-quality content residual rate after machine review, where the low-quality content residual rate after machine review is a proportion of low-quality content in each historical multimedia content that has passed machine review; Determining a target amount of manual review to be performed based on a target low-quality content residual rate, a target low-quality traffic residual rate, and the low-quality content residual rate after machine review, wherein the target low-quality traffic residual rate is a preset traffic ratio of low-quality content in each multimedia content; and the target low-quality content residual rate is a preset ratio of low-quality content in each multimedia content; Based on the flow rates of the multimedia contents, the target number of multimedia contents for manual review is determined among the multimedia contents as the multimedia contents to be manually reviewed.

2. The method according to claim 1, characterized in that The step of determining a target manual review amount to be performed based on the target low-quality content residual rate, the target low-quality traffic residual rate, and the low-quality content residual rate after machine review includes: Determining a first amount of multimedia content to be manually reviewed based on the low-quality content residual rate after machine review and the target low-quality content residual rate; Determining a second amount of multimedia content to be manually reviewed based on the low-quality content residual rate after machine review and the target low-quality traffic residual rate; The target manual review amount is determined based on a larger value of the first number and the second number.

3. The method according to claim 2, characterized in that The determining, based on the low-quality content residual rate after machine review and the target low-quality content residual rate, a first amount of multimedia content to be manually reviewed includes: Determining a proportion of multimedia content to be manually reviewed among the multimedia content based on the low-quality content residual rate after machine review and the target low-quality content residual rate; Obtaining the quantity of each multimedia content; Based on the quantity ratio and the quantity of each multimedia content, a first quantity of multimedia content to be manually reviewed is determined.

4. The method according to claim 2, characterized in that The determining, based on the low-quality content residual rate after machine review and the target low-quality traffic residual rate, a second amount of multimedia content to be manually reviewed includes: Determining a traffic proportion of multimedia content to be manually reviewed among the multimedia content based on the low-quality content residual rate after machine review and the target low-quality traffic residual rate; Obtaining the flow of each multimedia content; Based on the traffic ratio of the multimedia contents to be manually reviewed among the multimedia contents and the traffic of the multimedia contents, a second quantity of the multimedia contents to be manually reviewed is determined.

5. The method according to claim 1 or 2, characterized in that The step of determining the target manual review quantity of multimedia content from the multimedia content based on the respective traffic of the multimedia content as the multimedia content to be manually reviewed includes: The target manual review quantity of multimedia content is selected in descending order of the flow rate of each multimedia content as the multimedia content to be manually reviewed.

6. The method according to claim 5, characterized in that The step of selecting the target number of multimedia contents for manual review in descending order of the respective traffic volumes of the multimedia contents as the multimedia contents to be manually reviewed includes any one of the following: Determining multimedia content that has not been manually reviewed from the target number of multimedia contents for manual review, and using the multimedia content that has not been manually reviewed as the multimedia content to be manually reviewed; or Multimedia contents that have not been manually reviewed are determined from the multimedia contents, and the target number of multimedia contents that have not been manually reviewed are selected from the multimedia contents that have not been manually reviewed as the multimedia contents to be manually reviewed.

7. The method according to claim 1, characterized in that The low-quality content residual rate obtained after machine review includes: Acquire each historical multimedia content that has passed machine review, and perform manual review on each historical multimedia content to obtain historical manual review results; Based on the historical manual review results, the residual rate of low-quality content after machine review is determined.

8. The method according to claim 7, characterized in that The method further comprises: Based on a preset residual rate update cycle, the residual rate of the low-quality content after machine review is updated.

9. A multimedia content processing device, characterized in that: The device comprises: A first acquisition module is configured to acquire each multimedia content that has passed machine review and the flow rate of each multimedia content, wherein the flow rate of each multimedia content is updated according to a preset time period; A second acquisition module is configured to acquire a low-quality content residual rate after machine review, wherein the low-quality content residual rate after machine review is a proportion of low-quality content in each historical multimedia content that has passed machine review; a first determination module, configured to determine a target amount of manual review to be performed based on a target low-quality content residual rate and a target low-quality traffic residual rate, as well as the low-quality content residual rate after machine review, wherein the target low-quality traffic residual rate is a preset traffic ratio of low-quality content in each multimedia content; and the target low-quality content residual rate is a preset ratio of low-quality content in each multimedia content; The second determination module is configured to determine the target manual review quantity of multimedia content from among the multimedia content based on the respective traffic of the multimedia content, as the multimedia content to be manually reviewed.

10. An electronic device, characterized in that: The electronic device comprises: one or more processors; Memory; One or more computer programs, wherein the one or more computer programs are stored in the memory and configured to be executed by the one or more processors, and the one or more computer programs are configured to perform the method according to any one of claims 1 to 8.

Images