5G video service quality enhancement method, platform, equipment and medium

Abstract

The invention discloses a 5G video service quality enhancement method, platform, device and medium, and relates to the technical field of 5G services, the method comprises the following steps: calling historical service communication data, and mining a confrontation contradictory relationship of communication transmission; establishing a service enhancement module; generating a video interaction task, establishing interaction docking, and separating a network control plane from a data plane; interacting the 5G communication network state, performing transmission code rate self-adaption according to the confrontation contradictory relationship, and determining a transmission strategy; performing bidirectional redundancy compression on the pre-interaction video before and after transmission, and determining a data strategy; and fusing the transmission strategy and the data strategy, and performing 5G video service management. According to the invention, the technical problem of low quality and stability of the 5G video service caused by difficulty in balancing the bandwidth, the computing power and the delay in the existing 5G video service is solved, and the technical effects of realizing effective balance of the bandwidth, the computing power and the delay in the 5G video service and remarkably enhancing the quality and stability of the whole 5G video service are achieved.

Description

Technical Field

[0001] The present invention relates to the field of 5G service technology, and in particular to a 5G video service quality enhancement method, platform, device and medium. Background Art

[0002] With the large-scale commercial deployment of 5G technology, 5G video services have developed rapidly. However, there are still many technical challenges in its development process. In terms of video transmission, it is difficult to achieve an effective balance between bandwidth, computing power and latency. Although 5G networks have the characteristics of high bandwidth, they are still prone to bandwidth bottlenecks when facing massive video data transmission, especially high-definition, ultra-high-definition video and high-frame rate video. At the same time, the encoding and decoding of videos requires a lot of computing resources. If not handled properly, it will lead to increased latency, seriously affecting the real-time and smoothness of video services. For example, in some large-scale online video live broadcast activities, due to the large number of people watching online at the same time, the transmission and processing pressure of video data increases dramatically, and the screen often freezes and the latency is too high, which greatly reduces the user's viewing experience. There are also deficiencies in video redundancy processing technology. In high-concurrency service scenarios, there is a lack of effective service management strategies. When a large number of users request video services at the same time, network congestion and a sharp decline in service quality are prone to occur.

[0003] It is difficult to balance bandwidth, computing power, and latency in existing 5G video services, which leads to technical problems such as low quality and stability of 5G video services. Summary of the invention

[0004] The present application provides a 5G video service quality enhancement method, platform, device and medium, which are used to solve the technical problem that it is difficult to balance bandwidth, computing power and delay in existing 5G video services, thereby resulting in low quality and stability of 5G video services.

[0005] In view of the above problems, the present application provides a 5G video service quality enhancement method, platform, device and medium.

[0006] In a first aspect of the present application, a 5G video service quality enhancement method is provided, the method comprising: Call historical service communication data to mine the antagonistic and contradictory relationship of communication transmission, wherein the antagonistic and contradictory relationship is a relatively balanced relationship among bandwidth, computing power and delay; establish a service enhancement module, wherein the service enhancement module includes a bidirectional redundant branch and an AI generation block; generate a video interaction task and establish an interactive connection, separate the network control plane and the data plane, wherein the network control plane has the antagonistic and contradictory relationship built in, and the data plane has the service enhancement module built in; interact with the 5G communication network status, based on the network control plane, perform transmission bit rate adaptation for the video interaction task according to the antagonistic and contradictory relationship, and determine the transmission strategy; based on the data plane, in combination with the service enhancement module, perform bidirectional redundant compression on the pre-interaction video of the video interaction task before and after transmission, and determine the data strategy, wherein there is video fusion of the AI ​​generation task; integrate the transmission strategy and the data strategy, coordinate the network control plane and the data plane, and perform 5G video service management.

[0007] A second aspect of the present application provides a 5G video service quality enhancement platform, the platform comprising: A historical service communication data calling unit, the historical service communication data calling unit is used to call historical service communication data, and mine the antagonistic and contradictory relationship of communication transmission, wherein the antagonistic and contradictory relationship is a relatively balanced relationship between bandwidth, computing power and delay; a service enhancement module establishment unit, the service enhancement module establishment unit is used to establish a service enhancement module, wherein the service enhancement module includes a bidirectional redundant branch and an AI generation block; an interactive docking establishment unit, the interactive docking establishment unit is used to generate a video interaction task and establish an interactive docking, separate the network control plane and the data plane, wherein the network control plane has the antagonistic and contradictory relationship built in, and the data plane has the service enhancement module built in. Block; a transmission strategy determination unit, the transmission strategy determination unit is used to interact with the 5G communication network status, based on the network control plane, and adapt the transmission bit rate of the video interaction task according to the antagonistic and contradictory relationship to determine the transmission strategy; a data strategy determination unit, the data strategy determination unit is based on the data plane, combined with the service enhancement module, to perform bidirectional redundant compression on the pre-interaction video of the video interaction task before and after transmission, and determine the data strategy, wherein there is video fusion of AI generated tasks; a service management unit, the service management unit is used to integrate the transmission strategy and the data strategy, coordinate the network control plane and the data plane, and perform 5G video service management.

[0008] The third aspect of the present application provides an electronic device, which includes: a processor; a memory for storing instructions executable by the processor; wherein the processor is used to execute a 5G video service quality enhancement method provided in the present application.

[0009] The fourth aspect of the present application provides a computer-readable storage medium storing a computer program, which is used to execute a 5G video service quality enhancement method provided in the present application.

[0010] One or more technical solutions provided in this application have at least the following technical effects or advantages: Call historical service communication data to mine the antagonistic and contradictory relationship of communication transmission; establish a service enhancement module; generate video interaction tasks and establish interactive docking, separate the network control plane and the data plane; interact with the 5G communication network status, based on the network control plane, perform transmission rate adaptation according to the antagonistic and contradictory relationship, and determine the transmission strategy; based on the data plane, in combination with the service enhancement module, perform bidirectional redundant compression on the pre-interaction video before and after transmission, and determine the data strategy; integrate the transmission strategy and the data strategy, coordinate the network control plane and the data plane, and manage 5G video services. The technical effect of achieving an effective balance between bandwidth, computing power and latency in 5G video services and significantly enhancing the quality and stability of the overall 5G video services has been achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0012] Figure 1 A schematic diagram of a 5G video service quality enhancement method provided in an embodiment of the present application; Figure 2 A schematic diagram of the structure of a 5G video service quality enhancement platform provided in an embodiment of the present application.

[0013] Figure 3 A schematic diagram of the structure of an electronic device provided in this application.

[0014] Explanation of the accompanying drawings: historical service communication data calling unit 10, service enhancement module establishing unit 20, interactive docking establishing unit 30, transmission strategy determining unit 40, data strategy determining unit 50, service management unit 60, processor 21, memory 22, input device 23, output device 24. DETAILED DESCRIPTION

[0015] This application provides a 5G video service quality enhancement method, platform, device and medium to solve the technical problem that it is difficult to balance bandwidth, computing power and delay in existing 5G video services, resulting in low quality and stability of 5G video services.

[0016] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0017] Embodiment 1, as Figure 1 As shown, the present application provides a 5G video service quality enhancement method, the method comprising: Step S100: calling historical service communication data to mine the antagonistic and contradictory relationship of communication transmission, wherein the antagonistic and contradictory relationship is a relative balanced relationship among bandwidth, computing power and delay.

[0018] Specifically, we first need to accurately call historical service communication data from a massive historical database. These data come from a wide range of sources, including transmission records in various 5G video service scenarios, such as communication data in different time periods, different regions, different user groups, and different video types (HD, UHD, live, on-demand, etc.). The data covers computing power-related information such as the initial bandwidth allocation during each video transmission, the maximum and minimum bandwidth values ​​that can be dynamically allocated during the transmission process, the CPU usage of the server when processing video data, GPU acceleration, and the delay time from the video source to the successful reception of the data at the receiving end, as well as detailed information such as the corresponding network resource cost overhead. Advanced data mining algorithms and intelligent analysis models are used to deeply analyze these data. When mining the relationship between bandwidth, computing power and delay, statistical analysis and machine learning algorithms are used, such as multivariate linear regression analysis to initially explore the linear correlation trend between the three, and then deep neural networks are used to further fit the nonlinear relationship, and an accurate mathematical model is constructed to describe the intrinsic connection between them. When it is found that the bandwidth increases and the computing power increases, the delay of data transmission will show a significant downward trend. This is because wider bandwidth can transmit more data per unit time, and sufficient computing power can quickly encode, decode, compress, decompress and other processing operations on video data, reducing the waiting time of data during transmission. However, this improvement is not without cost. With the expansion of bandwidth and the increase of computing power, both the construction and maintenance costs of network infrastructure, as well as the energy consumption of servers, equipment purchase and upgrade costs, will increase accordingly. In the environment of full coverage of communication networks, an adaptive adjustment mechanism is established based on the confrontational and contradictory relationship that has been excavated. By deploying an intelligent monitoring and scheduling system at key nodes of the network, the bandwidth usage of the current network, the computing power load of each server, and the delay data of the ongoing video transmission are monitored in real time. Once it is found that the delay has a trend of increasing, according to the current cost budget and resource allocation, bandwidth resources are increased or more computing power is allocated to related tasks within the permitted range to quickly reduce the delay and ensure the smoothness of video services. When the network is idle or in a low-load state, the bandwidth and computing power usage will be appropriately reduced to save costs, achieve dynamic balance and optimization of the entire communication network under different business requirements and resource conditions, and lay a solid data foundation and decision-making basis for the subsequent improvement of 5G video service quality.

[0019] Step S200: Establishing a service enhancement module, wherein the service enhancement module includes a bidirectional redundancy branch and an AI generation block.

[0020] Specifically, a service enhancement module is established. This module is crucial to improving the quality of 5G video services. The bidirectional redundant branch is one of its core components, which utilizes multiple redundant characteristics in video data. Based on the temporal redundancy between adjacent frames, the similarity of the video in the temporal dimension is fully explored. By comparing the areas where the pixels between adjacent frames do not change much, the redundant information is accurately removed to achieve effective data compression. At the same time, based on the spatial redundancy between adjacent pixels, the law of image spatial distribution is deeply analyzed, and the areas with similar adjacent pixel values ​​or showing specific patterns are optimized to further reduce unnecessary data storage and transmission. Furthermore, supplemented by the statistical redundancy of the cells, the cells that appear frequently are efficiently encoded according to the probability distribution characteristics of the cells, reducing the overall complexity of the data. Based on these redundant elements, the principles of information theory are introduced. After repeated training with a large amount of sample data, a bidirectional redundant branch that can accurately perform video lossy compression is constructed, which significantly reduces the amount of data and improves transmission efficiency while ensuring the basic visual effects of the video. In addition, the service enhancement module also includes an AI generation block. With powerful artificial intelligence algorithms, this block can intelligently generate rich and diverse content based on the user's personalized needs, the characteristics of the video content, and the current service scenarios. For example, when watching sports event videos, it dynamically generates real-time score broadcasts, athlete data analysis, and other content based on the progress of the game, providing users with a richer and more in-depth viewing experience. The bidirectional redundant branch and the AI ​​generation block work together to inject powerful power into 5G video services, enabling them to achieve a qualitative leap in data processing and content innovation.

[0021] Step S300: Generate a video interaction task and establish an interactive connection, separate the network control plane and the data plane, wherein the network control plane has the antagonistic and contradictory relationship built in, and the data plane has the service enhancement module built in.

[0022] Specifically, first of all, according to the diverse video service needs of users and the preset functions of the system, various video interaction tasks are carefully generated. These tasks cover rich operations such as video playback, pause, fast forward, switching resolution, real-time interactive comments, and transmission and processing requirements of different types of videos (such as live broadcast, on-demand, video calls, etc.). Then, an accurate and efficient interactive docking mechanism is established to ensure seamless and real-time data interaction and command transmission between the video source, user terminal and each node of the network. On this basis, the entire network architecture is separated into the network control plane and the data plane. The network control plane has a precise built-in model of the antagonistic and contradictory relationship between bandwidth, computing power and delay that was previously discovered. It always monitors the overall operation status of the network, including the real-time occupancy of each link bandwidth, the allocation of computing power resources between different tasks, and the delayed data feedback during data transmission. Based on this information, through calculation and policy decision-making, key control parameters such as network resource allocation, routing selection and transmission protocol are dynamically adjusted to maintain the stability and efficiency of network services. The data plane is like a high-speed transportation channel of the network, which has a powerful service enhancement module built in. The bidirectional redundant branch in this module focuses on deep compression and optimization of video data before data transmission, removes various redundant information in the video data, effectively reduces the amount of data transmission, and improves transmission efficiency; at the same time, the AI ​​generation block intelligently generates rich additional information or special effects according to the video content and user needs during the data transmission process or at the receiving end, bringing users a more personalized and innovative video service experience. Through such network architecture separation and functional layout, a solid and flexible infrastructure has been laid for the all-round improvement of 5G video service quality.

[0023] Step S400: Interact with the 5G communication network status, based on the network control plane, and adapt the transmission bit rate of the video interaction task according to the antagonistic and contradictory relationship to determine the transmission strategy.

[0024] Specifically, it starts to interact with the 5G communication network status in a comprehensive and real-time manner. As the core decision-making unit, the network control plane performs adaptive adjustment of the transmission bit rate to determine the optimal transmission strategy based on the conflicting relationship between its built-in bandwidth, computing power and delay. It first accurately detects the real-time bandwidth of the network to understand the current bandwidth capacity available for data transmission, and closely monitors the allocation and remaining status of computing power resources, because sufficient computing power can support more complex video encoding and decoding operations, which in turn affects the setting of the transmission bit rate. In addition, delay feedback information during data transmission is collected, because delay is one of the key indicators for measuring the quality of video services. When the network bandwidth is relatively abundant and the computing power is sufficient, such as during low network usage periods or areas with strong network infrastructure performance, in order to provide users with the ultimate video viewing experience, the network control plane will increase the transmission bit rate so that the video can be transmitted at a higher resolution and smoother frame rate, showing a clear, delicate and coherent picture effect. On the contrary, if it is detected that bandwidth resources are tight, such as during network peak hours or in local network congestion areas, or computing power is limited due to multiple tasks running in parallel, in order to avoid video freezes or even interruptions, the network control plane will decisively reduce the transmission bit rate. At the same time, it may adjust the video encoding method and adopt a more efficient compression algorithm to minimize delays under limited bandwidth and computing power conditions, ensure that the video can be transmitted continuously and stably, and allow users to watch the video content normally. Therefore, under different network conditions, through this dynamic and intelligent transmission bit rate adaptation mechanism, the most appropriate transmission strategy can be determined to ensure the stability and smoothness of 5G video services.

[0025] Step S500: Based on the data plane and in combination with the service enhancement module, bidirectional redundant compression is performed on the pre-interaction video of the video interaction task before and after transmission to determine a data strategy, wherein there is video fusion of the AI ​​generated task.

[0026] Specifically, the data plane relies on its built-in service enhancement module to quickly start working for the pre-interaction video before transmission. The bidirectional redundancy branch in the service enhancement module is deeply explored to dig deep into the temporal redundancy between the adjacent frames in the video, accurately capture the similarities in the content of the adjacent frames, and efficiently compress these repeated information; carefully analyze the spatial redundancy between adjacent pixels, and reduce unnecessary pixel data storage according to the pixel distribution law; at the same time, make full use of the statistical redundancy of the information element, and optimize the information element encoding to reduce the amount of data. Through this series of complex and precise operations, the pre-compression processing of the pre-interaction video is realized, which effectively reduces the amount of data to be transmitted and reduces the burden on the subsequent transmission process. When the video is transmitted to the receiving end, the reverse function of the bidirectional redundancy branch, that is, the de-redundancy processing capability, is used again to accurately restore the redundant information in the compressed video according to the pre-set rules and algorithms, and perfectly reproduce the original picture quality and content of the pre-interaction video. In addition, if there is an AI generation task, the rich content generated by the AI ​​generation block will be deeply integrated with the video in this process. For example, if AI generates relevant special effects, text commentary or personalized recommendation information based on the video scene, it will seamlessly integrate these elements into the video screen or audio track at the right time and in an appropriate manner, thereby not only optimizing the transmission and restoration of video data, but also greatly enriching the content and expression of the video, and ultimately determining a complete data strategy to comprehensively improve the quality of 5G video services and user experience.

[0027] Step S600: Integrate the transmission strategy and the data strategy, coordinate the network control plane and the data plane, and perform 5G video service management.

[0028] Specifically, the transmission strategy and data strategy begin to be deeply integrated, and the network control plane and the data plane work together to jointly promote the management of 5G video services. The transmission strategy dominated by the network control plane focuses on network resource allocation and transmission link optimization. According to the dynamic balance of bandwidth, computing power and delay, it accurately controls key factors such as transmission bit rate and routing selection to ensure the efficient flow of video data in the network. The data strategy relied on by the data plane focuses on the processing and optimization of the video data itself, compressing and restoring the video with the help of bidirectional redundant branches, and integrating AI generation tasks to enrich the video content. The two functions are decoupled but interrelated. Through a special coordination mechanism, the network control plane transmits network status information to the data plane in real time so that it can flexibly adjust the data processing method according to the transmission conditions; the data plane reports the data processing requirements and feedback to the network control plane to assist it in optimizing the transmission strategy. In the entire 5G video service process, from the push of the video source to the reception and presentation of the user end, each link can be accurately managed and optimized to achieve a smooth and high-quality video playback experience, effectively improve the overall quality and stability of 5G video services, and meet the diverse needs of users for 5G video services in different scenarios.

[0029] In a possible implementation, step S600 further includes: Step S610: Setting a high concurrent service threshold.

[0030] Step S620: If the high concurrent service threshold is met, an intermediate service node is established, wherein the intermediate service node is an access network edge node of the receiving end.

[0031] Step S630: performing primary service management based on the transmission end and the intermediate service node, and performing secondary service management based on the intermediate service node and the receiving end.

[0032] Specifically, we first set a high concurrent service threshold, which is a key indicator determined based on a comprehensive analysis of the system's carrying capacity, network resource status, and past service data. It defines the upper limit of the number of video service requests that the system can effectively handle under normal operation. When the number of connections or data traffic requesting 5G video services at the same time exceeds this set value, it is considered to have entered a high concurrent service scenario.

[0033] Once it is detected that the current service request meets the set high concurrent service threshold, the process of establishing an intermediate service node is immediately started. The intermediate service node is positioned as the access network edge node at the receiving end, and its location and construction fully consider the network topology and user distribution. For example, setting up intermediate service nodes in various regional centers or large user gathering areas in a city can maximize the proximity to user terminals, effectively shorten the last mile of data transmission, and reduce signal transmission loss and delay.

[0034] Based on this network architecture adjustment, hierarchical service management is carried out. In the first service management stage, the transmission end first sends the video data to the intermediate service node. At this time, the intermediate service node assumes the functions of data aggregation, caching and preliminary processing. It can perform operations such as shunting and format conversion on the video stream from the transmission end. According to the network conditions and device performance of each receiving end, the video data is adaptively processed to avoid network jams or crashes caused by a large amount of data directly flowing to the receiving end. For example, for some low-bandwidth receiving ends, the intermediate service node can convert the video into a low-resolution version for transmission. In the second service management stage, the intermediate service node accurately distributes the preliminarily processed data to each receiving end. Since the intermediate service node has a deeper understanding and adaptation of the local network environment, it can adjust the transmission strategy in time according to the real-time feedback of the receiving end, such as dynamically adjusting the transmission bit rate or switching the transmission path when the local network is congested, thereby ensuring that each receiving end can obtain relatively stable and high-quality video services, effectively improving the effect and reliability of the entire 5G video service management in high-concurrency scenarios, and ensuring the user's viewing experience.

[0035] In a possible implementation, step S200 further includes: Step S210: taking the temporal redundancy between the previous and subsequent adjacent frames as the first redundancy element, taking the spatial redundancy between adjacent pixels as the second redundancy element, and taking the statistical redundancy of the cells as the third redundancy element.

[0036] Step S220: Based on the first redundant element, the second redundant element and the third redundant element as processing targets, information theory principles are introduced and sample training is used to construct a redundant branch, wherein the redundant branch is used to perform video lossy compression.

[0037] Specifically, the temporal redundancy between adjacent frames is established as the first redundant element. As a continuous sequence of images, adjacent frames of a video usually have a high degree of similarity. For example, in a video showing a calm lake, the ripples on the lake and the outlines of the surrounding scenery remain almost unchanged in multiple consecutive frames. This large amount of repeated image information in the temporal dimension constitutes the available temporal redundancy. Secondly, the spatial redundancy between adjacent pixels plays an important role as the second redundant element. An image is composed of many pixels. In a local area, adjacent pixels often show similarities or gradients in attributes such as color and brightness. For example, in an image with a blue sky background, the blue hue and brightness of adjacent pixels change slightly and continuously. By identifying and processing this spatial pixel correlation, the amount of data storage can be effectively reduced. Finally, the statistical redundancy of cells is regarded as the third redundant element. At the cell level involved in video data transmission, the probability of occurrence of different cells varies. Some cells appear frequently, while others appear less frequently. Based on this statistical law, high-frequency cells can be specially encoded to reduce the redundancy of the overall data. These three redundant elements together provide the core basis and data processing direction for the subsequent construction of efficient redundant branches.

[0038] Based on the above three redundant elements, the redundant branch is constructed. The principle of information theory is introduced. Information theory provides a theoretical foundation for data compression, which can measure the amount of information and redundancy in data. A large number of video samples are used as training materials. During the training process, the entropy coding and other technical means in information theory are used. For temporal redundancy, the difference between adjacent frames is calculated and encoded through the inter-frame prediction algorithm to remove duplicate information; for spatial redundancy, the discrete cosine transform method is used to convert the image pixel block to the frequency domain, highlight its energy concentration characteristics and then quantize and encode; for the statistical redundancy of the signal element, variable length coding operations such as Huffman coding are performed according to the probability of the signal element. Through such sample training and principle application, a redundant branch capable of performing lossy video compression is gradually constructed. Under the premise of sacrificing a small amount of video quality, the storage and transmission volume of video data is significantly reduced, the transmission efficiency and resource utilization of 5G video services are improved, and a solid data processing foundation is laid for the subsequent enhancement of video service quality.

[0039] In a possible implementation, step S220 further includes: Step S221: invert the redundant branch to serve as a de-redundant branch.

[0040] Step S222: establishing a trigger rule for the receiving end of the redundant branch and a trigger rule for the transmitting end of the de-redundant branch.

[0041] Specifically, in order to restore the video data processed by the redundant branch at the receiving end of video transmission, the redundant branch needs to be inverted to construct a redundancy removal branch. First, a series of algorithms and operation processes used by the redundant branch in processing videos are deeply analyzed. In the temporal redundancy processing of adjacent frames, the inter-frame prediction algorithm is used to record the motion vector and residual information between frames to achieve compression; for the spatial redundancy between adjacent pixels, discrete cosine transform, quantization and entropy coding are used to transform and compress pixel data; in terms of statistical redundancy of cells, Huffman coding operations are implemented based on the probability distribution of cell occurrence. Then, the redundancy removal branch is constructed with the inverse process of these algorithms as the core. For temporal redundancy, according to the stored motion vector and residual information, the received compressed video frames are gradually restored to the original video frame sequence through reverse motion compensation and residual addition operations; for spatial redundancy, the inverse quantization operation is first performed, and then the inverse discrete cosine transform is used to convert the frequency domain data back to the spatial domain pixel data; and for the statistical redundancy of cells, the decoding operation is performed according to the Huffman coding table to restore the encoded data to the original cell sequence. Through the integration and application of this series of reverse operation algorithms for different redundancy types, a redundant branch was successfully constructed, which enables the video data to be effectively restored when the video is transmitted to the receiving end, ensuring that users receive high-quality and complete video content.

[0042] Constructing the trigger rules of the receiving end of the redundant branch and the trigger rules of the transmitting end of the de-redundant branch is the key link to ensure efficient and accurate processing of video data. For the trigger rules of the receiving end of the redundant branch, when the video source is ready and sends a transmission request to the network control plane, the network control plane will comprehensively evaluate multiple factors such as the current network bandwidth, computing resources, and the overall transmission task queue. If it is determined that the network has the ability to handle new video transmission tasks and there is no conflict with higher priority tasks, it will send an instruction to start the redundant branch to the receiving end, and at the same time pass the relevant video source information and network parameters. After receiving the instruction and information, the receiving end immediately starts the redundant branch and starts to compress the incoming video data based on time, space, and cell statistical redundancy to optimize the data volume for subsequent transmission. For the trigger rules of the transmitting end of the de-redundant branch, when the compressed video data arrives at the network buffer of the receiving end and undergoes preliminary integrity and legality verification, the receiving end will feedback the receiving status information to the network control plane. The network control plane determines whether it is appropriate to start the de-redundant branch for decompression based on this information and the current local network conditions. If the conditions are met, the command to start the redundancy removal branch is sent to the receiving end, along with the key parameters required for decompression, such as encoding table information, motion vector data, etc. After receiving the command, the receiving end immediately activates the redundancy removal branch, decompresses the compressed video data according to the established reverse algorithm, and gradually restores the original video data, thereby ensuring that the video can be effectively processed and restored according to reasonable rules before and after transmission, ensuring the smoothness and stability of the entire video service process.

[0043] In a possible implementation, step S500 further includes: Step S510: Based on the redundant branch, pre-process the pre-interaction video to determine a compressed video.

[0044] Step S520: executing communication transmission of the compressed video according to the transmission strategy, and performing redundancy processing of the compressed video based on the de-redundancy branch at the receiving end to restore the pre-interaction video.

[0045] Specifically, the previously constructed redundancy branch is fully utilized to pre-process the pre-interaction video. The redundancy branch performs in-depth analysis and optimization of the pre-interaction video based on its accurate processing capabilities for temporal redundancy between adjacent frames, spatial redundancy between adjacent pixels, and statistical redundancy of cells. Through the inter-frame prediction and motion compensation algorithm for temporal redundancy, a large amount of similar image information between adjacent frames is identified and removed, and only the key inter-frame change data is retained; for spatial redundancy, the discrete cosine transform technology is used to convert the image pixel data to the frequency domain for quantization and encoding processing, effectively reducing the amount of redundant data between adjacent pixels; in terms of statistical redundancy of cells, Huffman coding and other operations are implemented according to the probability distribution of the occurrence of cells to further compress the data. After this series of complex and efficient processing processes, the original pre-interaction video is converted into a compressed video with a greatly reduced data volume, creating favorable conditions for subsequent communication transmission.

[0046] The communication transmission of compressed video is carried out strictly in accordance with the determined transmission strategy. The transmission strategy is formulated based on multiple factors such as network bandwidth, computing power and delay monitored by the network control plane. It clarifies the key elements such as the most suitable transmission bit rate, routing selection and transmission protocol in the current network environment. During the transmission process, the compressed video data is pushed stably from the source to the receiving end according to these strategies. When the compressed video reaches the receiving end, redundancy processing is performed based on the pre-set redundancy removal branch. The redundancy removal branch uses the reverse algorithm corresponding to the redundancy branch to perform reverse motion compensation and residual addition operations for temporal redundancy using the stored motion vector and residual information to restore the original inter-frame relationship; for spatial redundancy, first perform inverse quantization and then convert the frequency domain data back to spatial domain pixel data through inverse discrete cosine transform; for cell statistical redundancy, perform decoding operations based on the Huffman coding table. Through this series of de-redundancy processing steps, the compressed video is gradually restored to the original pre-interactive video, ensuring that the receiving user can watch video content that is basically consistent with the source. This ensures the quality of video services while effectively improving the transmission efficiency and resource utilization of video data in the 5G network environment.

[0047] In a possible implementation, step S500 further includes: Step S530: Identify the pre-interaction video and determine multimedia elements.

[0048] Step S540: traverse the multimedia elements, extract the first multimedia element for data processing, and determine the first processing result.

[0049] Step S550: mapping the first processing result to the pre-interaction video, and integrating and determining the processing result of the multimedia content in the pre-interaction video.

[0050] Specifically, the pre-interaction video is comprehensively and meticulously identified to accurately determine the various multimedia elements contained therein. Through video analysis tools and algorithms, the multi-dimensional information contained in the video is deeply excavated. First, focus on the image part of the video to identify the main elements in the picture, such as people, scenery, objects, etc. At the same time, analyze the basic attributes of the image such as resolution, color mode, encoding method, etc., and further analyze the changes in the image in the time dimension, such as the movement trajectory of objects in the picture, the switching rhythm of scenes, etc., because these elements have a key mapping relationship with the time dimension in subsequent processing, which can help us better locate and restore the processing effect. In addition to images, detailed identification of voice elements is also carried out to determine the number of voice channels, sampling frequency, voice content, and the start and end time of voice. Attention is also paid to the distribution of voice on the time axis to clarify its coordination relationship with the video screen. In addition, other multimedia elements such as subtitles, video effects, hyperlinks, etc. are also identified one by one, and their respective presentation forms, appearance timing, and their relationship with the overall video in time and space are clarified. The entire recognition process aims to build a complete and clear multimedia element framework. For example, when processing one of the items, such as an image, the processed image result can be accurately located, mapped and extracted in the original video based on the previously determined mapping relationship such as the time dimension. In this way, the accuracy of the processing can be guaranteed, and the processing efficiency can be greatly improved, thus avoiding ineffective work and laying a solid foundation for further optimizing the overall quality of the pre-interaction video.

[0051] Starting from the first multimedia element, each element is checked and screened in turn. When the first multimedia element is located, the corresponding data processing operation is carried out according to its type. If the first multimedia element is a specific area in the video image, such as a facial image of a person, an image recognition and processing algorithm is used to first detect the edge of the area and accurately outline the facial contour. Then, the skin color correction technology is used to adjust the facial skin color to make it more natural. At the same time, the image sharpening algorithm is used to highlight the details of the facial features, such as the clarity of the eyes and the texture of the lips. After this series of image processing, the optimized facial image of the person is obtained as the first processing result. If the first multimedia element is a voice narration in the audio track, the voice processing technology is used to first perform voice noise reduction processing to remove background noise interference, and then perform voice enhancement to improve the volume and clarity of the voice. Special effects processing such as voice speed change and pitch change can also be performed according to needs, and finally a processed audio clip is formed as the first processing result. If the first multimedia element is subtitle text, a text processing algorithm is used to check and correct grammatical errors in the subtitles. If multi-language requirements are involved, machine translation technology can be used to translate the subtitles into a specified language and adjust the display format of the subtitles, such as font color, background transparency, etc., so as to determine the processed subtitle text as the first processing result, thereby making full preparations for the subsequent mapping and integration in the pre-interactive video.

[0052] After the first processing result is obtained, precise mapping and integration operations are performed within the original framework of the pre-interactive video. If the first processing result is optimized data about the image, such as the image has been color enhanced, details of a specific area have been enlarged or blurred, etc., the processed image data is accurately replaced with the original image data according to the time code and spatial coordinate information of the image in the pre-interactive video, ensuring that when the video is played, the frame sequence where the image is located can naturally and smoothly present the processed visual effect, and maintain coherence with the transition between the previous and next frames. If the first processing result is an adjustment in audio, such as volume standardization, audio special effects addition, etc., the processed audio clip will be seamlessly integrated into the original audio track according to the starting and ending points of the audio on the timeline, so that the audio and video screen are synchronized and coordinated to avoid the phenomenon of sound and screen misalignment. For the processing results of other multimedia elements such as subtitles, the modified subtitle text, style or special effects are also completely mapped back to the corresponding position in the original video according to their time position and display layout in the pre-interactive video. By such detailed mapping and integration operations on all the first processing results, the overall processing results of the multimedia content in the pre-interaction video are finally determined, so that the video is optimized and improved in all multimedia dimensions, bringing users a better and richer video viewing experience.

[0053] In a possible implementation, step S520 further includes: Step S521: If there is an AI generated task, semantic recognition and analysis are performed on the AI ​​generated task to determine the generated elements.

[0054] Step S522: traverse the generation elements, call and combine the standardized generation elements based on the generation component library, perform intelligent generation processing on the personalized elements, and determine the generation results. The generation component library contains a standardized component set.

[0055] Step S523: Merge the pre-interaction video and the generated result as display content of the receiving end.

[0056] Specifically, when an AI-generated task is detected, the semantic recognition and parsing process based on the deep learning framework is started. First, the text related to the task is input into the neural network model, which adopts an architecture that combines a convolutional neural network (CNN) and a recurrent neural network (RNN). CNN is responsible for extracting local features in the text, such as keywords, phrase structures, etc., and converting the text into feature vectors through the operations of the convolution layer and the pooling layer. RNN performs sequence modeling on the feature vectors, using the long short-term memory network (LSTM) unit to capture the semantic dependencies of the text, analyze the contextual associations between words, and the overall semantic logic of the sentence. In the model training stage, a large amount of annotated task text data is used for supervised learning, and the weight parameters of the neural network are continuously adjusted to improve the accuracy of semantic recognition. When the AI-generated task text to be parsed is input, the model outputs a semantic feature representation, and then determines the generation elements based on the predefined semantic templates and rule base. For example, if the task text involves a description of a specific scene, the model can identify key objects, environmental attributes, emotional tendencies and other elements in the scene, and convert them into structured data forms, such as an array of objects stored in JSON format, where each object contains information such as the element name, attribute value, and related constraints, providing accurate and computer-processable data basis for subsequent element generation processing, ensuring that AI generation tasks can accurately meet video service needs and effectively drive subsequent intelligent content generation processes.

[0057] Start the traversal operation of the determined generation elements. For the parts that are determined to be standardized generation elements, the task is completed with the help of a generation component library that is pre-built and stores a rich set of standardized components. The standardized component set in this component library covers a variety of types, such as general video format conversion plug-ins and standardized video filter modules in video processing; common audio format adaptation components and basic audio special effects units in the audio field; standard font style templates and general text typesetting rule sets in the text category. According to the specific requirements description of the generation elements, the corresponding standardized components are quickly located and called by accurately matching the index information in the component library, and then these called components are organically spliced ​​and configured according to the preset combination logic and rules to ensure that they work together in terms of function and data interaction. For those parts that are identified as personalized elements, the intelligent generation processing flow is started. Using an intelligent generation model built based on a deep learning algorithm and a generative adversarial network (GAN), it has been fully trained on large-scale multimodal data. It can generate personalized content that is unique, innovative, and highly suitable for the task context based on the specific semantic information, style requirements, and related features of the pre-interaction video carried by the generated elements through complex neural network calculation and reasoning processes. For example, it can create a unique soundtrack melody for a specific video scene, or generate customized animation effects that are closely related to the video plot and artistically appealing. Finally, the results of the combination of standardized components are summarized and integrated with the personalized generated content to determine the complete generation results, so as to be fully prepared for the subsequent integration with the pre-interaction video, so as to realize the intelligent and personalized enrichment and expansion of video content and enhance the overall video service experience.

[0058] Start the fusion process of the pre-interactive video and the generated results, aiming to construct the final display content of the receiving end. For the text information in the generated results, such as commentary, subtitles, etc., according to the time code and video screen layout rules, the text is accurately superimposed on the corresponding frame of the pre-interactive video to ensure that the text display position is reasonable, clear and readable, and complements the main content of the screen. The font, color, size and other style attributes of the text are adjusted synchronously to adapt to the video style and visual effects. If the generated results contain audio elements, such as background music, special effects sound, etc., use audio mixing technology to finely set the volume ratio, channel balance, fade-in and fade-out effects and other parameters according to the role positioning of the audio and the characteristics of the original audio track of the pre-interactive video, so that the newly added audio and the original video audio are naturally integrated to create a harmonious, unified and layered auditory environment. As for the visual effects or animation clips in the generated results, with the help of video synthesis algorithms, according to the scope of the effects or animations, triggering time and matching requirements with the video scene, they are seamlessly embedded into the specific frame sequence or scene of the pre-interactive video, ensuring that during the video playback, the visual effects are presented smoothly and naturally, with smooth transitions, and closely matching the overall rhythm and picture changes of the video. Through the all-round and refined integration of the generated results of text, audio, vision and other aspects, the receiving end display content is finally generated, bringing users a rich, diverse and immersive video viewing experience, effectively improving the quality of video services.

[0059] Embodiment 2, based on the same inventive concept as a method for enhancing 5G video service quality in the aforementioned embodiment, Figure 2 As shown, the present application provides a 5G video service quality enhancement platform, and the platform and method embodiments in the embodiments of the present application are based on the same inventive concept. Among them, the platform includes: The historical service communication data calling unit 10 is used to call historical service communication data and mine the antagonistic and contradictory relationship of communication transmission, wherein the antagonistic and contradictory relationship is a relative balance relationship among bandwidth, computing power and delay.

[0060] The service enhancement module establishment unit 20 is used to establish a service enhancement module, wherein the service enhancement module includes a bidirectional redundant branch and an AI generation block.

[0061] The interactive docking establishment unit 30 is used to generate a video interactive task and establish an interactive docking, separate the network control plane and the data plane, wherein the network control plane has the antagonistic and contradictory relationship built in, and the data plane has the service enhancement module built in.

[0062] The transmission strategy determination unit 40 is used to interact with the 5G communication network status, based on the network control plane, and to adapt the transmission bit rate of the video interaction task according to the antagonistic and contradictory relationship to determine the transmission strategy.

[0063] A data strategy determination unit 50, based on the data plane and in combination with the service enhancement module, performs bidirectional redundant compression on the pre-interaction video of the video interaction task before and after transmission to determine the data strategy, wherein there is video fusion of the AI ​​generated task.

[0064] The service management unit 60 is used to integrate the transmission strategy and the data strategy, coordinate the network control plane and the data plane, and perform 5G video service management.

[0065] Furthermore, the service management unit 60 further includes: A high concurrent service threshold setting unit, wherein the high concurrent service threshold setting unit is used to set a high concurrent service threshold.

[0066] An intermediate service node establishment unit, the intermediate service node establishment unit is used to establish an intermediate service node if the high concurrent service threshold is met, wherein the intermediate service node is an access network edge node of the receiving end.

[0067] A primary service management unit performs primary service management based on the transmission end and the intermediate service node, and performs secondary service management based on the intermediate service node and the receiving end.

[0068] Furthermore, the service enhancement module establishing unit 20 further includes: The redundant element determination unit is used to use the temporal redundancy between the adjacent frames as the first redundant element, the spatial redundancy between the adjacent pixels as the second redundant element, and the statistical redundancy of the information element as the third redundant element.

[0069] A redundant branch construction unit, wherein the redundant branch construction unit takes the first redundant element, the second redundant element and the third redundant element as processing targets, introduces information theory principles and uses sample training as a method to construct redundant branches, wherein the redundant branches are used to perform video lossy compression.

[0070] Furthermore, the redundant branch construction unit also includes: A redundant branch acquisition unit is used to invert the redundant branch to serve as a de-redundant branch.

[0071] A receiving end trigger rule establishing unit is used to establish a receiving end trigger rule for the redundant branch and a transmitting end trigger rule for the de-redundant branch.

[0072] Furthermore, the data strategy determination unit 50 further includes: A compressed video determining unit, wherein the compressed video determining unit performs pre-processing on the pre-interaction video based on the redundant branch to determine a compressed video.

[0073] A pre-interaction video restoration unit, the pre-interaction video restoration unit is used to execute the communication transmission of the compressed video according to the transmission strategy, and at the receiving end, perform redundancy processing on the compressed video based on the de-redundancy branch to restore the pre-interaction video.

[0074] Furthermore, the data strategy determination unit 50 further includes: A multimedia element determination unit is used to identify the pre-interaction video and determine multimedia elements.

[0075] The first processing result determination unit is used to traverse the multimedia elements, extract the first multimedia elements for data processing, and determine the first processing result.

[0076] A pre-interaction video mapping unit is used to map the first processing result to the pre-interaction video, and integrate and determine the processing result of the multimedia content in the pre-interaction video.

[0077] Furthermore, the pre-interaction video restoration unit further includes: A generating element determining unit, wherein the generating element determining unit is used to perform semantic recognition and analysis on the AI ​​generating task if there is one, and determine the generating elements.

[0078] A generation result determination unit is used to traverse the generation elements, call and combine the standardized generation elements based on the generation component library, perform intelligent generation processing on the personalized elements, and determine the generation results. The generation component library contains a standardized component set.

[0079] A display content acquisition unit is used to merge the pre-interaction video and the generated result as display content of the receiving end.

[0080] Embodiment 3, Figure 3 The schematic diagram of the structure of the electronic device provided in the third embodiment of the present invention shows a block diagram of an exemplary electronic device suitable for implementing the implementation mode of the present invention. Figure 3The electronic device shown is only an example and should not limit the functions and scope of use of the embodiments of the present invention. Figure 3 As shown, the electronic device includes a processor 21, a memory 22, an input device 23 and an output device 24; the number of processors 21 in the electronic device can be one or more. Figure 3 Taking a processor 21 as an example, the processor 21, memory 22, input device 23 and output device 24 in the electronic device can be connected through a bus or other means. Figure 3 The example of connecting through bus is taken in the following.

[0081] The memory 22, as a computer-readable storage medium, can be used to store software programs, computer executable programs and modules, such as program instructions / modules corresponding to a 5G video service quality enhancement method in an embodiment of the present application. The processor 21 executes various functional applications and data processing of the computer device by running the software programs, instructions and modules stored in the memory 22, that is, implementing the above-mentioned 5G video service quality enhancement method.

[0082] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0083] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

[0084] This specification and the drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. A 5G video service quality enhancement method, characterized in that: The method comprises: Call historical service communication data to mine the antagonistic and contradictory relationship of communication transmission, where the antagonistic and contradictory relationship is the relative balance relationship between bandwidth, computing power and latency; Establishing a service enhancement module, wherein the service enhancement module includes a bidirectional redundancy branch and an AI generation block; Generate a video interaction task and establish an interactive connection, separate a network control plane and a data plane, wherein the network control plane has the antagonistic and contradictory relationship built in, and the data plane has the service enhancement module built in; Interact with the 5G communication network status, based on the network control plane, and adaptively transmit the bit rate of the video interaction task according to the confrontation and contradiction relationship to determine the transmission strategy; Based on the data plane, in combination with the service enhancement module, the pre-interaction video of the video interaction task is subjected to bidirectional redundancy compression before and after transmission, and a data strategy is determined, wherein there is video fusion of the AI ​​generation task; The transmission strategy and the data strategy are integrated, and the network control plane and the data plane are coordinated to perform 5G video service management.

2. A 5G video service quality enhancement method as claimed in claim 1, characterized in that: The 5G video service management includes: Set a high concurrent service threshold; If the high concurrent service threshold is met, establishing an intermediate service node, wherein the intermediate service node is an access network edge node of the receiving end; A primary service management is performed based on the transmission end and the intermediate service node, and a secondary service management is performed based on the intermediate service node and the receiving end.

3. A 5G video service quality enhancement method as claimed in claim 1, characterized in that: The service enhancement module includes a bidirectional redundant branch, including: The temporal redundancy between the previous and next adjacent frames is the first redundancy factor, the spatial redundancy between adjacent pixels is the second redundancy factor, and the statistical redundancy of the information element is the third redundancy factor; Based on the first redundant element, the second redundant element and the third redundant element as processing targets, information theory principles are introduced and sample training is used to construct a redundant branch, wherein the redundant branch is used to perform video lossy compression.

4. A 5G video service quality enhancement method as claimed in claim 3, characterized in that: The method further comprises: Inverting the redundant branch to serve as a de-redundant branch; A triggering rule for the receiving end of the redundant branch and a triggering rule for the transmitting end of the de-redundant branch are established.

5. A 5G video service quality enhancement method as claimed in claim 4, characterized in that: Bidirectional redundancy compression of pre-interactive video before and after transmission, including: Based on the redundant branch, pre-processing the pre-interaction video to determine a compressed video; The communication transmission of the compressed video is performed according to the transmission strategy, and at the receiving end, redundancy processing of the compressed video is performed based on the de-redundancy branch to restore the pre-interaction video.

6. A 5G video service quality enhancement method as claimed in claim 5, characterized in that: Bidirectional redundancy compression of pre-interactive video before and after transmission, including: Identify the pre-interaction video and determine multimedia elements; Traversing the multimedia elements, extracting a first multimedia element for data processing, and determining a first processing result; The first processing result is mapped to the pre-interaction video, and the processing result of the multimedia content in the pre-interaction video is integrated and determined.

7. A 5G video service quality enhancement method as claimed in claim 5, characterized in that: After restoring the pre-interaction video, the method includes: If there is an AI generated task, perform semantic recognition and analysis on the AI ​​generated task to determine the generated elements; Traversing the generation elements, calling and combining the standardized generation elements based on the generation component library, performing intelligent generation processing on the personalized elements, and determining the generation results, wherein the generation component library includes a standardized component set; The pre-interaction video and the generated result are integrated as display content of the receiving end.

8. A 5G video service quality enhancement platform, characterized in that: The platform is used to execute a 5G video service quality enhancement method according to any one of claims 1 to 7, and the platform includes: A historical service communication data calling unit, the historical service communication data calling unit is used to call historical service communication data to mine the antagonistic and contradictory relationship of communication transmission, wherein the antagonistic and contradictory relationship is a relative balance relationship between bandwidth, computing power and delay; A service enhancement module establishment unit, the service enhancement module establishment unit is used to establish a service enhancement module, wherein the service enhancement module includes a bidirectional redundant branch and an AI generation block; An interactive docking establishment unit, the interactive docking establishment unit is used to generate a video interactive task and establish an interactive docking, separate a network control plane and a data plane, wherein the network control plane has the antagonistic and contradictory relationship built in, and the data plane has the service enhancement module built in; A transmission strategy determination unit, the transmission strategy determination unit is used to interact with the 5G communication network status, based on the network control plane, and perform transmission bit rate adaptation on the video interaction task according to the antagonistic and contradictory relationship to determine the transmission strategy; A data strategy determination unit, wherein the data strategy determination unit performs bidirectional redundancy compression on the pre-interaction video of the video interaction task before and after transmission based on the data plane and in combination with the service enhancement module to determine a data strategy, wherein there is video fusion of an AI generated task; A service management unit, wherein the service management unit is used to integrate the transmission strategy and the data strategy, coordinate the network control plane and the data plane, and perform 5G video service management.

9. An electronic device, characterized in that: The electronic device comprises: processor; a memory for storing instructions executable by the processor; The processor is used to execute a 5G video service quality enhancement method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, so that the computer program is used to execute a 5G video service quality enhancement method as described in any one of claims 1 to 7.

Images