Digital Human Video Generation Method and Related Equipment Based on Ascend AI Processor

Abstract

This invention provides a method and related equipment for generating digital human videos based on the Ascend AI processor, optimizing digital human synthesis efficiency through dynamic resource scheduling and task splitting mechanisms. The method includes: after a user submits a task, a task forwarding service performs sensitive word detection and generates synthesized audio; the main video synthesis service dynamically adjusts its deployment based on Ascend AI processor resources, splitting the task into sub-tasks and distributing them to sub-services; frame integrity verification ensures synthesis quality, and finally, the video is generated and stored. This invention significantly improves resource utilization and task execution reliability, making it suitable for large-scale digital human video generation scenarios.

Description

Technical Field

[0001] This invention relates to the field of programming technology, and in particular to a method and related equipment for generating digital human videos based on the Ascend AI processor. Background Technology

[0002] With the rapid development of digital technology, the application of digital humans (or virtual humans) is gradually becoming widespread in various fields, covering industries such as entertainment, education, healthcare, and customer service. Digital humans can accurately simulate human appearance, behavior, and emotions, making the interaction between humans and computers more natural and fluid, and improving interaction efficiency. This trend has not only opened up innovative business models for enterprises but also brought entirely new experiences to users. Especially against the backdrop of the continuous maturation of virtual reality (VR) and augmented reality (AR) technologies, the application of digital human technology is becoming increasingly widespread, and its potential in areas such as digital marketing, social media, and online education has attracted increasing attention and importance from organizations.

[0003] The importance of concurrent execution of AI tasks is closely related to the large video memory of the Ascend AI processor. With the increasing demands of AI applications, concurrent task execution has become crucial for improving computing efficiency and accelerating data processing. The Ascend AI processor, with its large-capacity video memory, can deploy multiple services simultaneously on a single GPU, optimizing resource utilization and reducing waiting time between tasks. This allows for not only improved overall system performance when handling multiple AI tasks but also ensures efficient parallel execution of different tasks without sacrificing computing power. In this way, enterprises can enhance business processing capabilities, shorten response times, and better meet the rapid demands of the market and customers. Summary of the Invention

[0004] The technical problem to be solved by this invention is to propose a digital human video generation method and related equipment based on the Ascend AI processor, which aims to solve the problems of concurrent task processing and resource optimization management in the digital human video generation process.

[0005] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a digital human video generation method based on the Ascend AI processor, comprising the following steps:

[0006] S10. The user terminal submits a digital human synthesis task, including selecting a digital human model, inputting synthesized text, or uploading audio.

[0007] S20: The task forwarding service performs sensitive word detection on the input content and generates synthesized audio through the speech synthesis service;

[0008] S30. Send the digital human model parameters and synthesized audio to the main video compositing service;

[0009] S40, based on the Ascend AI processor, dynamically adjusts the deployment method of the main video synthesis service according to the available video memory and quantity;

[0010] S50: The main video compositing service is split into several sub-tasks and distributed to the sub-video compositing services.

[0011] S60. The sub-video compositing service executes sub-tasks and returns results, while the main video compositing service verifies the integrity of the composite frame count.

[0012] S70: After compressing the composite video, upload it to the storage server and update the task status.

[0013] Furthermore, the sensitive word detection in step S20 includes:

[0014] S21. When a user uploads audio, sensitive words are detected after the audio is automatically converted into text by speech recognition.

[0015] S22. When the user inputs synthesized text, perform sensitive word detection directly;

[0016] S23. After the detection is passed, the synthesized audio is generated.

[0017] Furthermore, the dynamic adjustment described in step S40 includes:

[0018] Based on the available video memory size and number of processors of the Ascend AI processor in the server, the main video compositing service is dynamically deployed, and several sub-video compositing services are associated under each main service.

[0019] Furthermore, the task splitting described in step S50 includes:

[0020] The main video compositing service obtains the number of currently available sub-services, splits the total task into the number of currently available sub-services + 1 sub-task, assigns the first number of sub-tasks to the sub-services, and executes the last sub-task by the main service.

[0021] Furthermore, the frame count integrity verification in step S60 includes:

[0022] The main video compositing service counts the number of composite frames returned by the subtasks. If the sum of these counts equals the target total number of frames, the task is considered successful.

[0023] Furthermore, the subtask execution described in step S60 includes:

[0024] The sub-video compositing service calculates the required number of frames to be composited based on the audio duration and video frame rate, and then returns the generated video frames or images to the main service.

[0025] Furthermore, the synthetic video compression described in step S70 includes:

[0026] The main video compositing service integrates the compositing results of all subtasks by frame sequence and uploads them to cloud storage or a file server via network protocols.

[0027] The present invention also provides a digital human video generation device based on the Ascend AI processor, comprising:

[0028] The task submission module is used by user terminals to submit digital human synthesis tasks, including selecting a digital human model, inputting synthetic text, or uploading audio.

[0029] The audio synthesis module is used by the task forwarding service to detect sensitive words in the input content and generate synthesized audio through the speech synthesis service.

[0030] The material sending module is used to send the digital human model parameters and synthesized audio to the main video compositing service;

[0031] The dynamic adjustment module is used to dynamically adjust the deployment method of the main video compositing service based on the available and quantity of GPU memory of the Ascend AI processor;

[0032] The task splitting module is used to split the main video compositing service into several sub-tasks and distribute them to the sub-video compositing services;

[0033] The service verification module is used for the sub-video compositing service to execute sub-tasks and return results, and for the main video compositing service to verify the integrity of the number of composite frames.

[0034] The video compression module is used to compress the synthesized video, upload it to the storage server, and update the task status.

[0035] The present invention also provides a computer device, the computer device including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the digital human video generation method based on the Ascend AI processor as described in any of the preceding claims.

[0036] The present invention also provides a storage medium storing a computer program, which, when executed by a processor, can implement the digital human video generation method based on the Ascend AI processor as described in any of the preceding claims.

[0037] The beneficial effects of this invention are as follows: Through the technical solution of this invention, tasks can be effectively split and allocated, resource utilization optimized, and the execution efficiency of digital human synthesis tasks improved, with the collaborative work of multiple main video synthesis services and sub-video synthesis services. Furthermore, by comparing the number of frames in the synthesis result with the target number of frames, the accuracy and completeness of the digital human synthesis task are ensured. The implementation of this solution significantly improves the resource scheduling efficiency and task completion reliability during the synthesis process. Attached Figure Description

[0038] The specific structure of the present invention will now be described in detail with reference to the accompanying drawings.

[0039] Figure 1 This is an application scenario diagram of the digital human video generation method based on the Ascend AI processor according to an embodiment of the present invention;

[0040] Figure 2 This is a flowchart of a digital human video generation method based on the Ascend AI processor according to an embodiment of the present invention;

[0041] Figure 3 This is a flowchart of the sensitive word detection process according to an embodiment of the present invention;

[0042] Figure 4 This is a block diagram of a digital human video generation device based on the Ascend AI processor according to an embodiment of the present invention;

[0043] Figure 5 This is a schematic block diagram of a computer device according to an embodiment of the present invention. Detailed Implementation

[0044] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0045] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0046] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0047] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0048] like Figure 1 As shown, the digital human video generation method based on the Ascend AI processor provided in this embodiment of the invention is applied in applications such as... Figure 1In this application environment, the terminal communicates with the server via a network. The user terminal submits a digital human synthesis task, including selecting a digital human model, inputting synthetic text, or uploading audio; the task forwarding service performs sensitive word detection on the input content and generates synthetic audio through the speech synthesis service; the digital human model parameters and synthetic audio are sent to the main video synthesis service; the server dynamically adjusts the deployment method of the main video synthesis service based on the available and quantity of GPU memory of the Ascend AI processor; the main video synthesis service splits the task into several sub-tasks and distributes them to the sub-video synthesis services; the sub-video synthesis services execute the sub-tasks and return the results, and the main video synthesis service verifies the integrity of the synthesized frame count; the synthesized video is compressed and uploaded to the storage server, and the task status is updated. Through the technical solution of this invention, with the collaborative work of multiple main video synthesis services and sub-video synthesis services, tasks can be effectively split and allocated, resource utilization can be optimized, and the execution efficiency of digital human synthesis tasks can be improved. Furthermore, by comparing the number of frames in the synthesis result with the target number of frames, the accuracy and integrity of the digital human synthesis task are ensured. The implementation of this solution significantly improves the resource scheduling efficiency and task completion reliability in the synthesis process. The terminal can be, but is not limited to, various personal computers, laptops, smartphones, tablets, and portable wearable devices. The server can be implemented using a standalone server or a server cluster consisting of multiple servers. The invention will now be described in detail through specific embodiments.

[0049] like Figure 2 As shown, an embodiment of the present invention is: a digital human video generation method based on the Ascend AI processor, comprising the following steps:

[0050] S10. The user terminal submits a digital human synthesis task, including selecting a digital human model, inputting synthetic text, or uploading audio.

[0051] S20: The task forwarding service performs sensitive word detection on the input content and generates synthesized audio through the speech synthesis service.

[0052] like Figure 3 As shown, in one specific embodiment, the sensitive word detection in step S20 includes:

[0053] S21. When a user uploads audio, sensitive words are detected after the audio is automatically converted into text by speech recognition.

[0054] S22. When the user inputs synthesized text, perform sensitive word detection directly;

[0055] S23. After the detection is passed, the synthesized audio is generated.

[0056] In this embodiment, when a user uploads audio, the task forwarding service uses automatic speech recognition (ASR) to convert the audio into text and performs sensitive word detection on the converted text. If sensitive words are detected, the digital human synthesis task fails to execute; otherwise, the detection passes and proceeds to the next step.

[0057] When a user inputs synthetic text, the task forwarding service performs sensitive word detection on the text. If a sensitive word is detected, the digital human synthesis task fails to execute; otherwise, the detection passes and proceeds to the next step.

[0058] If the detection passes, the synthesized text or the converted text is sent to an idle speech synthesis service for synthesis to obtain synthesized audio.

[0059] S30. Send the digital human model parameters and synthesized audio to the main video synthesis service.

[0060] In this embodiment, the task forwarding service sends the digital human model parameters selected by the user and the user-uploaded audio or synthesized audio that has passed the detection to the currently idle main video synthesis service.

[0061] S40, based on the Ascend AI processor, dynamically adjusts the deployment method of the main video compositing service according to the available video memory and quantity.

[0062] In one specific embodiment, the dynamic adjustment in step S40 includes:

[0063] Based on the available video memory size and number of processors of the Ascend AI processor in the server, the main video compositing service is dynamically deployed, and several sub-video compositing services are associated under each main service.

[0064] In this embodiment, the deployment method of the main video synthesis service is dynamically adjusted according to the available video memory size and the number of Ascend AI processors on the server, based on the actual business scenario, to ensure reasonable resource allocation.

[0065] S50: The main video compositing service is split into several sub-tasks and distributed to the sub-video compositing services.

[0066] In one specific embodiment, the task splitting in step S50 includes:

[0067] The main video compositing service obtains the number of currently available sub-services, splits the total task into the number of currently available sub-services + 1 sub-task, assigns the first number of sub-tasks to the sub-services, and executes the last sub-task by the main service.

[0068] In this embodiment, after the idle main video compositing service receives the digital human compositing task, the status of the main video compositing service changes to "busy" and the status of the digital human compositing task in the database is updated to "in execution".

[0069] The main video compositing service determines the duration of the user-uploaded audio or synthesized audio and the frame rate of the target synthesized video based on the detected audio. Calculate the total number of frames required for synthesis. :

[0070]

[0071] The total number of frames in the target synthesized video is known. Enter the start frame number of the video. End frame number The final generated frame sequence .

[0072] The final synthesized video frame sequence The generation is based on the following steps and mathematical relationships:

[0073] Forward sequence From arrive ,Right now:

[0074] ;

[0075] This is a collection A sequence of elements.

[0076] Reverse sequence From arrive ,

[0077] ;

[0078] This is a collection A sequence of elements.

[0079] By combining the forward and reverse sequences, the initial frame sequence is obtained. :

[0080] ;

[0081] If the length of the generated frame sequence is less than the given total number of frames Then continue splicing forward and reverse sequences until the total length is at least equal to .

[0082] After each stitch, a new frame sequence is created. length It becomes:

[0083] ;

[0084] until Finally, if the length of T exceeds Then, the excess part is cut off to ensure that the total length is exactly 1 / 3. .

[0085] ;

[0086] The final generated frame sequence The following conditions must be met:

[0087] This is the initial frame sequence;

[0088] ;

[0089] Number of splices time Then cut off the excess part;

[0090] ;

[0091] The main video compositing service communicates with the sub-video compositing services via a health interface to obtain the number of currently available sub-video compositing services, and dynamically updates the final frame sequence calculated in the digital human compositing task based on the actual situation. The process is divided into several sub-digital human synthesis tasks. The number of sub-digital human synthesis tasks that are currently available is sent to the sub-video synthesis interface, and the last sub-digital human synthesis task is sent to the main video synthesis interface.

[0092] The number of sub-video compositing interfaces and main video compositing interfaces is: :

[0093] ;

[0094] The basic length of each subframe sequence is:

[0095] ;

[0096] in yes Divide by The integer part represents the minimum length of each subframe sequence.

[0097] Remaining elements yes Divide by Remainder:

[0098] ;

[0099] The remaining elements are evenly distributed to the front. In a list of sublists, make these sublists one element longer than the other sublists.

[0100] when At that time, the first The length of each subframe sequence is:

[0101] ;

[0102] when At that time, the first The length of each subframe sequence is:

[0103] ;

[0104] The elements of each subframe sequence are from Extract from.

[0105] For the The starting position of each subframe sequence for:

[0106] ;

[0107] End position of subframe sequence ;

[0108] Each sub-video compositing service that receives the sub-digital human compositing task executes the task and notifies the main video compositing service of the compositing result after the task is completed.

[0109] S60. The sub-video compositing service executes the sub-task and returns the result, while the main video compositing service verifies the integrity of the composite frame count.

[0110] In one specific embodiment, the frame number integrity verification in step S60 includes:

[0111] The main video compositing service counts the number of composite frames returned by the subtasks. If the sum of these counts equals the target total number of frames, the task is considered successful.

[0112] In one specific embodiment, the subtask execution in step S60 includes:

[0113] The sub-video compositing service calculates the required number of frames to be composited based on the audio duration and video frame rate, and then returns the generated video frames or images to the main service.

[0114] In this embodiment, the main video synthesis service checks the synthesis result and determines whether the digital human synthesis task is successful based on the sum of the number of synthesized images or the number of sub-digital human synthesized video frames and the total number of frames to be synthesized.

[0115] S70: After compressing the composite video, upload it to the storage server and update the task status.

[0116] In one specific embodiment, the synthetic video compression in step S70 includes:

[0117] The main video compositing service integrates the compositing results of all subtasks by frame sequence and uploads them to cloud storage or a file server via network protocols.

[0118] In this embodiment, when the synthesis result of the sub-video synthesis service is consistent with the total number of frames of the target synthesized video, the main video synthesis service considers the digital human synthesis task to be successfully executed, compresses the synthesized video and uploads it to the file storage server, updates the status of the digital human synthesis task in the database to "success", and the status of the main video synthesis service changes to "idle".

[0119] The technical effects of the technical solution in this application include:

[0120] By rationally allocating tasks and dynamically scheduling resources, the efficiency of digital human video generation has been significantly improved. This efficiency improvement is not only reflected in the speed of video generation but also in the full utilization of resources, avoiding waste of hardware resources.

[0121] Effective task management and resource scheduling enhance system stability and reliability. Dynamic resource allocation strategies ensure stable system operation even under high loads, reducing the risk of service interruptions due to resource contention.

[0122] By ensuring high-quality synthesis results, the user experience is enhanced. Result verification technology ensures high accuracy and completeness in video synthesis, allowing users to obtain video content that meets their expectations, thereby increasing user satisfaction with the digital human video generation system.

[0123] The system design incorporates scalability, enabling it to expand as user needs grow. This scalability means the system can adapt to different business scales and changing requirements without frequent hardware replacements or large-scale system upgrades.

[0124] like Figure 4 As shown, the present invention also provides a digital human video generation device based on the Ascend AI processor, comprising:

[0125] The task submission module 10 is used by the user terminal to submit digital human synthesis tasks, including selecting a digital human model, inputting synthetic text, or uploading audio.

[0126] The audio synthesis module 20 is used by the task forwarding service to detect sensitive words in the input content and generate synthesized audio through the speech synthesis service.

[0127] The material sending module 30 is used to send the digital human model parameters and synthesized audio to the main video compositing service;

[0128] The dynamic adjustment module 40 is used to dynamically adjust the deployment method of the main video synthesis service based on the available and quantity of GPU memory of the Ascend AI processor;

[0129] The task splitting module 50 is used to split the main video compositing service into several sub-tasks and distribute them to the sub-video compositing services;

[0130] Service verification module 60 is used for the sub-video synthesis service to execute sub-tasks and return results, and for the main video synthesis service to verify the integrity of the synthesized frame count.

[0131] The video compression module 70 is used to compress the synthesized video, upload it to the storage server, and update the task status.

[0132] It should be noted that those skilled in the art can clearly understand that the specific implementation process of the above-mentioned digital human video generation device based on the Ascend AI processor can be referred to the corresponding description in the foregoing method embodiments. For the sake of convenience and brevity, it will not be repeated here.

[0133] The aforementioned digital human video generation device based on the Ascend AI processor can be implemented as a computer program, which can, for example... Figure 5 It runs on the computer device shown.

[0134] Please see Figure 5 , Figure 5 This is a schematic block diagram of a computer device provided in an embodiment of this application. The computer device 500 can be a terminal or a server. The terminal can be an electronic device with communication functions, such as a smartphone, tablet, laptop, desktop computer, personal digital assistant, or wearable device. The server can be a standalone server or a server cluster composed of multiple servers.

[0135] See Figure 5 The computer device 500 includes a processor 502, a memory, and a network interface 505 connected via a system bus 501. The memory may include a non-volatile storage medium 503 and internal memory 504.

[0136] The non-volatile storage medium 503 may store an operating system 5031 and a computer program 5032. The computer program 5032 includes program instructions that, when executed, cause the processor 502 to perform a digital human video generation method based on the Ascend AI processor.

[0137] The processor 502 provides computing and control capabilities to support the operation of the entire computer device 500.

[0138] The internal memory 504 provides an environment for the operation of the computer program 5032 in the non-volatile storage medium 503. When the computer program 5032 is executed by the processor 502, the processor 502 can execute a digital human video generation method based on the Ascend AI processor.

[0139] This network interface 505 is used for network communication with other devices. Those skilled in the art will understand that... Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device 500 to which the present application is applied. The specific computer device 500 may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0140] The processor 502 is used to run a computer program 5032 stored in the memory to implement the digital human video generation method based on the Ascend AI processor as described above.

[0141] It should be understood that in the embodiments of this application, the processor 502 may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0142] It will be understood by those skilled in the art that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program includes program instructions and can be stored in a storage medium, which is a computer-readable storage medium. The program instructions are executed by at least one processor in the computer system to implement the process steps of the embodiments of the above methods.

[0143] Therefore, the present invention also provides a storage medium. This storage medium can be a computer-readable storage medium. The storage medium stores a computer program, wherein the computer program includes program instructions. When executed by a processor, the program instructions cause the processor to perform the digital human video generation method based on the Ascend AI processor as described above.

[0144] The storage medium can be any computer-readable storage medium capable of storing program code, such as a USB flash drive, portable hard drive, read-only memory (ROM), magnetic disk, or optical disk.

[0145] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0146] In the several embodiments provided by this invention, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For example, the division of each unit is merely a logical functional division, and there may be other division methods in actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed.

[0147] The steps in the method of this invention can be adjusted, merged, or reduced in order according to actual needs. The units in the device of this invention can be merged, divided, or reduced according to actual needs. Furthermore, the functional units in the various embodiments of this invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0148] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a terminal, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention.

[0149] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for generating digital human videos based on the Ascend AI processor, characterized in that, Includes the following steps: S10. The user terminal submits a digital human synthesis task, including selecting a digital human model, inputting synthesized text, or uploading audio. S20: The task forwarding service performs sensitive word detection on the input content and generates synthesized audio through the speech synthesis service; S30. Send the digital human model parameters and synthesized audio to the main video compositing service; S40. Based on the available video memory size and number of processors of the Ascend AI processor in the server, dynamically deploy the main video compositing service, with several sub-video compositing services associated under each main service; S50: The main video compositing service obtains the number of currently available sub-services, splits the total task into the number of currently available sub-services + 1 sub-task, assigns the first number of sub-tasks to the sub-services, and executes the last sub-task by the main service. Step S50 specifically includes: The main video compositing service determines the duration of the user-uploaded audio or synthesized audio that has passed the detection process and the frame rate of the target synthesized video. Calculate the total number of frames required for synthesis. : ； The total number of frames in the target synthesized video is known. Enter the start frame number of the video. End frame number The final generated frame sequence; The final synthesized video frame sequence The generation is based on the following steps and mathematical relationships: Forward sequence From arrive ,Right now: ； This is a collection A sequence of elements; Reverse sequence From arrive , ； This is a collection A sequence of elements; By combining the forward and reverse sequences, the initial frame sequence is obtained. : ； If the length of the generated frame sequence is less than the given total number of frames Then continue splicing forward and reverse sequences until the total length is at least equal to ; After each splicing, a new frame sequence is generated. length It becomes: ； until Finally, if the length of T exceeds Then cut off the excess part to ensure that the total length is exactly 1 / 3. ; ； The final generated frame sequence The following conditions must be met: This is the initial frame sequence; ； Number of splices time Then cut off the excess part; ； The main video compositing service communicates with the sub-video compositing services via a health interface to obtain the number of currently available sub-video compositing services, and dynamically updates the final frame sequence calculated in the digital human compositing task based on the actual situation. The process is divided into several sub-digital human synthesis tasks. The number of digital human synthesis sub-tasks that are currently available in the sub-services is sent to the sub-video synthesis interface. The last digital human synthesis sub-task is sent to the main video synthesis interface. The number of sub-video compositing interfaces and main video compositing interfaces is: : ； The basic length of each subframe sequence is: ； in yes Divide by The integer part represents the minimum length of each subframe sequence; Remaining elements yes Divide by Remainder: ； The remaining elements are evenly distributed to the front. In a list of sublists, find sublists whose length is one more element than the other sublists; when At that time, the first The length of each subframe sequence is: ； when At that time, the first The length of each subframe sequence is: ； The elements of each subframe sequence are from Extract from; For the The starting position of each subframe sequence for: ； End position of subframe sequence ; Each sub-video compositing service that receives the sub-digital human compositing task executes the task and notifies the main video compositing service of the compositing result after the task is completed; S60. The main video compositing service counts the number of composite frames returned by the subtask. If the sum of these frames equals the target total number of frames, the task is considered successful. S70: After compressing the composite video, upload it to the storage server and update the task status.

2. The digital human video generation method based on the Ascend AI processor according to claim 1, characterized in that, The sensitive word detection in step S20 includes: S21. When a user uploads audio, it is automatically converted into text through speech recognition and then sensitive word detection is performed. S22. When the user inputs synthesized text, perform sensitive word detection directly; S23. After the detection is passed, the synthesized audio is generated.

3. The digital human video generation method based on the Ascend AI processor according to claim 1, characterized in that, The subtask execution in step S60 includes: The sub-video compositing service calculates the required number of frames to be composited based on the audio duration and video frame rate, and then returns the generated video frames or images to the main service.

4. The digital human video generation method based on the Ascend AI processor according to claim 1, characterized in that, The synthetic video compression described in step S70 includes: The main video compositing service integrates the compositing results of all subtasks by frame sequence and uploads them to cloud storage or a file server via network protocols.

5. A digital human video generation device based on the Ascend AI processor, characterized in that, include: The task submission module is used by user terminals to submit digital human synthesis tasks, including selecting a digital human model, inputting synthetic text, or uploading audio. The audio synthesis module is used by the task forwarding service to detect sensitive words in the input content and generate synthesized audio through the speech synthesis service. The material sending module is used to send the digital human model parameters and synthesized audio to the main video compositing service; The dynamic adjustment module is used to dynamically deploy the main video compositing service based on the available video memory size and number of processors of the Ascend AI processor in the server, with several sub-video compositing services associated under each main service; The task splitting module is used by the main video compositing service to obtain the number of currently available sub-services, split the total task into the number of currently available sub-services + 1 sub-task, allocate the first number of sub-tasks to the sub-services, and execute the last sub-task by the main service; The task splitting module is specifically used by the main video compositing service to determine the duration of the user-uploaded audio or synthesized audio that has passed the detection test and the frame rate of the target synthesized video. Calculate the total number of frames required for synthesis. : ； The total number of frames in the target synthesized video is known. Enter the start frame number of the video. End frame number The final generated frame sequence ; The final synthesized video frame sequence The generation is based on the following steps and mathematical relationships: Forward sequence From arrive ,Right now: ； This is a collection A sequence of elements; Reverse sequence From arrive , ； This is a collection A sequence of elements; By combining the forward and reverse sequences, the initial frame sequence is obtained. : ； If the length of the generated frame sequence is less than the given total number of frames Then continue splicing forward and reverse sequences until the total length is at least equal to ; After each splicing, a new frame sequence is generated. length It becomes: ； until Finally, if the length of T exceeds Then cut off the excess part to ensure that the total length is exactly 1 / 3. ; ； The final generated frame sequence The following conditions must be met: This is the initial frame sequence; ； Number of splices time Then cut off the excess part; ； The main video compositing service communicates with the sub-video compositing services via a health interface to obtain the number of currently available sub-video compositing services, and dynamically updates the final frame sequence calculated in the digital human compositing task based on the actual situation. The process is divided into several sub-digital human synthesis tasks. The number of digital human synthesis sub-tasks that are currently available in the sub-services is sent to the sub-video synthesis interface. The last digital human synthesis sub-task is sent to the main video synthesis interface. The number of sub-video compositing interfaces and main video compositing interfaces is: : ； The basic length of each subframe sequence is: ； in yes Divide by The integer part represents the minimum length of each subframe sequence; Remaining elements yes Divide by Remainder: ； The remaining elements are evenly distributed to the front. In a list of sublists, find sublists whose length is one more element than the other sublists; when At that time, the first The length of each subframe sequence is: ； when At that time, the first The length of each subframe sequence is: ； The elements of each subframe sequence are from Extract from; For the The starting position of each subframe sequence for: ； End position of subframe sequence ; Each sub-video compositing service that receives the sub-digital human compositing task executes the task and notifies the main video compositing service of the compositing result after the task is completed; The service verification module is used by the main video compositing service to count the number of composite frames returned by subtasks. If the sum of these frames equals the target total number of frames, the task is considered successful. The video compression module is used to compress the synthesized video, upload it to the storage server, and update the task status.

6. A computer device, characterized in that: The computer device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements the digital human video generation method based on the Ascend AI processor as described in any one of claims 1 to 4.

7. A storage medium, characterized in that: The storage medium stores a computer program, which, when executed by a processor, can implement the digital human video generation method based on the Ascend AI processor as described in any one of claims 1 to 4.

Images