Video editing system and method based on content identification

Abstract

The invention relates to the technical field of video editing, discloses a video editing system and method based on content recognition, and aims to solve the problems that traditional video editing software is high in hardware requirement, complex in operation, not suitable for a mobile terminal and the like. The task distribution module selects a proper computing node for online video editing for the to-be-edited video according to the video format and the video size, so that the system editing efficiency is improved, the requirements on local software and hardware are reduced, meanwhile, high-efficiency video editing processing is realized by utilizing powerful computing power resources of the cloud server, and then the video editing module is used for editing the video. And finally, screening out the video clips meeting the requirements based on a confidence evaluation module, screening out the video clips meeting the requirements, and splicing the screened video clips together according to a time sequence to complete video clip.

Description

Technical Field

[0001] The present invention relates to the technical field of video editing, and in particular to a video editing system and method based on content recognition. Background Art

[0002] With the popularization of the Internet and the rapid development of multimedia technology, video has become an indispensable part of people's daily life. In many video application scenarios, such as short video platforms, when users share wonderful video highlights, they often need to edit the complete video before publishing it. However, most existing video editing software relies on high-performance hardware devices, which have certain requirements for the configuration of hardware devices, and the operation is complicated, making it difficult to run smoothly on devices with limited computing power such as mobile terminals. In addition, traditional video editing methods often lead to a large waste of computing power and cannot meet the growing demand for video editing.

[0003] Online video editing is a convenient way to edit videos. It allows users to perform cloud video editing directly through an Internet browser without installing additional software. This improves the efficiency of video editing, especially in today's world where the demand for rapid iteration and content sharing is growing. Most of the existing online video editing systems have the following problems: in the process of online cloud editing of videos, there is a lack of judgment on the node status of cloud editing of videos. If the selected editing node does not meet the editing requirements of the video to be edited, it will not only affect the editing efficiency and quality of the video, but also affect the node itself. Therefore, the present invention proposes a video editing system and method based on content recognition. Summary of the invention

[0004] The purpose of the present invention is to provide a video editing system and method based on content recognition to solve the above technical problems.

[0005] A video editing system based on content recognition, the system comprising: an editing task acceptance module, an editing task allocation module, a video editing module and a confidence evaluation module;

[0006] The editing task accepting module is used to import the video to be edited and obtain the basic data of the video to be edited;

[0007] The editing task allocation module is used to allocate the editing nodes of the videos to be edited according to the basic data of each video to be edited acquired by the editing task acceptance module;

[0008] The video editing module is used to edit the video based on the video content;

[0009] The confidence evaluation module is used to evaluate the confidence of the clipped video clips and output the target video.

[0010] As a further description of the solution of the present invention, the working process of the editing task accepting module includes:

[0011] The editor opens the online video editing client page in the browser, uploads the video file to be edited to the system and temporarily stores it in the media library;

[0012] Get the size and content of each video file to be edited.

[0013] As a further description of the solution of the present invention, the working process of the editing task allocation module includes:

[0014] The system obtains the size and format of the video file to be edited;

[0015] Screen out the editing nodes that meet the requirements according to the size and format of the video file to be edited. If there are idle editing nodes among the editing nodes that meet the requirements, the editor manually selects any editing node for online video editing;

[0016] If there is no idle node among the editing nodes that meet the requirements, the system will send online editing requests to all the editing nodes that meet the requirements in turn. When the first editing node that meets the requirements receives the online editing request, it will determine whether it is idle. If so, it will add the current video file to the editing list and perform online editing. If not, it will transfer the current editing task to the next editing node that meets the requirements.

[0017] The next editing node that meets the requirements determines whether it is in an idle state. If so, it adds the current video file to the editing list for online editing. Otherwise, it continues to transfer the current editing task to the next editing node that meets the requirements.

[0018] The editing task is passed in turn among all the editing nodes that meet the requirements until the current video file is added to the editing list by one of the editing nodes that meets the requirements.

[0019] As a further description of the solution of the present invention, if there is no idle node among the editing nodes that meet the requirements, and the online editing request at the system has not been responded to within a set time period, the current online editing request is discarded and a new online editing request is reissued.

[0020] As a further description of the solution of the present invention, the working process of screening out the clipping nodes that meet the requirements includes:

[0021] Obtaining the operating parameter data of all trusted clipping nodes;

[0022] The matching degree K of all trusted clip nodes and the current video file is calculated by the following formula:

[0023] K = αA*βB;

[0024] In the formula, α and β are weight coefficients, A is the format matching index, and B is the memory matching index;

[0025] The matching degree K between the trusted clipping node and the current video file is compared with the set matching degree standard value K0. If the matching degree K between the trusted clipping node and the current video file is greater than the set matching degree standard value K0, it means that the trusted node meets the requirements, otherwise, the trusted node does not meet the requirements.

[0026] As a further description of the scheme of the present invention, the format matching index A is related to the format of the video file to be edited. If the format of the video file to be edited is within the video format range processed by the editing node, A=1; if the format of the video file to be edited is not within the video format range processed by the editing node, A=0.

[0027] As a further description of the scheme of the present invention, the memory matching index B is related to the size of the video file to be edited. The size M of the video file to be edited is obtained. If the memory space available to the editing node is greater than 2M, B=1; if the memory space available to the editing node is less than or equal to 2M and greater than M, B=0.5; if the memory space available to the editing node is less than or equal to M, B=0.

[0028] As a further description of the solution of the present invention, the working process of the video editing module includes:

[0029] Obtain video content, and extract n video clips based on specific scenes in the video content;

[0030] Number each video clip in the following format: where i belongs to n, is the initial moment of the i-th video clip, is the end time of the i-th video clip, σ i is the sequence number of the i-th video clip.

[0031] As a further description of the solution of the present invention, the working process of the confidence assessment module includes:

[0032] According to the initial and end times of the n video clips, the n videos are sequentially input into the trained action recognition network model, the confidence evaluation result of each video is output, and each video clip is renumbered in the following format: Among them, P i Give each a confidence score;

[0033] The editors re-screen the video clips that do not meet the confidence score requirements, and select the misjudged video clips to add to the video clips that meet the confidence score requirements;

[0034] According to the numbering format, the initial and end times of all video clips that meet the confidence score requirements are obtained, and all video clips are merged in chronological order to complete the online video editing.

[0035] A video editing method based on content recognition, the method comprising the following steps:

[0036] Step S1: The editor uploads the video to be edited to the online video editing client page;

[0037] Step S2: The system obtains the size and format of the video file to be edited;

[0038] Step S3, filtering out editing nodes that meet the requirements for online editing of the video to be edited according to the content obtained in step S2;

[0039] Step S4, selecting a qualified editing node according to the current state of the editing node that satisfies the online editing of the video to be edited;

[0040] Step S5, selecting the video clips that meet the requirements according to the scenes in the video content;

[0041] Step S6: Confidence scoring is performed on all the video clips that meet the requirements by using the trained model;

[0042] Step S7: select the final video clips according to the scoring results, splice them in chronological order, and complete the video editing.

[0043] Beneficial effects of the present invention:

[0044] The present invention aims to solve the problems of traditional video editing software having high hardware requirements, complex operations, and being unsuitable for mobile terminals. Through the editing task acceptance module, the video to be edited is uploaded to the online cloud editing, and the task allocation module selects a suitable computing node for online video editing for the video to be edited according to the video format and video size, thereby improving the editing efficiency of the system and reducing the demand for local software and hardware. At the same time, the powerful computing resources of the cloud server are utilized to realize efficient video editing processing, and then, the target video clips are edited through the video editing module. Finally, the video clips that meet the requirements are screened out based on the confidence evaluation module, and the screened out video clips that meet the requirements are spliced ​​together in chronological order to complete the video editing. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The present invention will be further described below in conjunction with the accompanying drawings.

[0046] Figure 1 It is a partial structural diagram of the video editing system based on content recognition provided by the present invention. DETAILED DESCRIPTION

[0047] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] See also Figure 1 As shown, the present invention is a video editing system based on content recognition, the system comprising: an editing task acceptance module, an editing task allocation module, a video editing module and a confidence evaluation module;

[0049] The editing task accepting module is used to import the video to be edited and obtain the basic data of the video to be edited;

[0050] The editing task allocation module is used to allocate the editing nodes of the videos to be edited according to the basic data of each video to be edited acquired by the editing task acceptance module;

[0051] The video editing module is used to edit the video based on the video content;

[0052] The confidence evaluation module is used to evaluate the confidence of the clipped video clips and output the target video.

[0053] Through the above-mentioned calculation scheme, the present invention aims to solve the problems of traditional video editing software having high hardware requirements, complex operations, and not being suitable for mobile terminals. Through the editing task acceptance module, the video to be edited is uploaded to the online cloud editing, and the task allocation module selects a suitable computing node for online video editing for the video to be edited according to the video format and video size, thereby improving the system editing efficiency and reducing the demand for local software and hardware. At the same time, the powerful computing power resources of the cloud server are utilized to realize efficient video editing processing, and then, the target video clips are edited through the video editing module. Finally, the video clips that meet the requirements are screened out based on the confidence evaluation module, and the screened video clips that meet the requirements are spliced ​​together in chronological order to complete the video editing.

[0054] The working process of the editing task accepting module includes:

[0055] The editor opens the online video editing client page in the browser, uploads the video file to be edited to the system and temporarily stores it in the media library;

[0056] Get the size and content of each video file to be edited.

[0057] The working process of the editing task allocation module includes:

[0058] The system obtains the size and format of the video file to be edited;

[0059] Screen out the editing nodes that meet the requirements according to the size and format of the video file to be edited. If there are idle editing nodes among the editing nodes that meet the requirements, the editor manually selects any editing node for online video editing;

[0060] If there is no idle node among the editing nodes that meet the requirements, the system will send online editing requests to all the editing nodes that meet the requirements in turn. When the first editing node that meets the requirements receives the online editing request, it will determine whether it is idle. If so, it will add the current video file to the editing list and perform online editing. If not, it will transfer the current editing task to the next editing node that meets the requirements.

[0061] The next editing node that meets the requirements determines whether it is in an idle state. If so, it adds the current video file to the editing list for online editing. Otherwise, it continues to transfer the current editing task to the next editing node that meets the requirements.

[0062] The editing task is passed in turn among all the editing nodes that meet the requirements until the current video file is added to the editing list by one of the editing nodes that meets the requirements.

[0063] Through the above technical solution, when this embodiment allocates nodes, first, nodes that meet the requirements are screened out. If there are idle nodes among the nodes that meet the requirements, the editor manually selects any editing node to perform online video editing. If there are no idle nodes among the nodes that meet the requirements, the system sends online editing requests to all editing nodes that meet the requirements in turn. The editing task is passed in turn to all editing nodes that meet the requirements until the current video file is added to the editing list by one of the editing nodes that meet the requirements. Each editing request has a life cycle. If it is not responded to within the life cycle, it will be discarded immediately.

[0064] If there is no idle node among the editing nodes that meet the requirements, and the online editing request at the system has not been responded to within the set time period, the current online editing request will be discarded and a new online editing request will be reissued.

[0065] The working process of screening out the clipping nodes that meet the requirements includes:

[0066] Obtaining the operating parameter data of all trusted clipping nodes;

[0067] The matching degree K of all trusted clip nodes and the current video file is calculated by the following formula:

[0068] K = αA*βB;

[0069] In the formula, α and β are weight coefficients, A is the format matching index, and B is the memory matching index;

[0070] The matching degree K between the trusted clipping node and the current video file is compared with the set matching degree standard value K0. If the matching degree K between the trusted clipping node and the current video file is greater than the set matching degree standard value K0, it means that the trusted node meets the requirements, otherwise, the trusted node does not meet the requirements.

[0071] The format matching index A is related to the format of the video file to be edited. If the format of the video file to be edited is within the video format range processed by the editing node, A=1; if the format of the video file to be edited is not within the video format range processed by the editing node, A=0.

[0072] The memory matching index B is related to the size of the video file to be edited. The size M of the video file to be edited is obtained. If the memory space available to the editing node is greater than 2M, B=1; if the memory space available to the editing node is less than or equal to 2M and greater than M, B=0.5; if the memory space available to the editing node is less than or equal to M, B=0.

[0073] Through the above technical solution, in order to screen out the clipping nodes that meet the requirements, the formula K=αA*βB is used to calculate the matching degree of all the trusted clipping nodes and the current video file, and then the format matching index is assigned according to the video format content. If the format of the video file to be clipped is within the video format range processed by the clipping node, the format matching index is 1, if the format of the video file to be clipped is not within the video format range processed by the clipping node, the format matching index is 0, and the memory matching index is assigned according to the video size to obtain the size M of the video file to be clipped. If the memory space available for the clipping node is greater than 2M, the memory matching index is 1, if the memory space available for the clipping node is less than or equal to 2M and greater than M, the memory matching index is 0.5, if the memory space available for the clipping node is less than or equal to M, the memory matching index is 0, and finally the matching degree K of the trusted clipping node and the current video file is compared with the set matching degree standard value K0. If the matching degree K of the trusted clipping node and the current video file is greater than the set matching degree standard value K0, it means that the trusted node meets the requirements, otherwise, the trusted node does not meet the requirements.

[0074] The working process of the video editing module includes:

[0075] Obtain video content, and extract n video clips based on specific scenes in the video content;

[0076] Number each video clip in the following format: where i belongs to n, is the initial moment of the i-th video clip, is the end time of the i-th video clip, σ i is the sequence number of the i-th video clip.

[0077] The working process of the confidence assessment module includes:

[0078] According to the initial and end times of the n video clips, the n videos are sequentially input into the trained action recognition network model, the confidence evaluation result of each video is output, and each video clip is renumbered in the following format: Among them, P i Give each a confidence score;

[0079] The editors re-screen the video clips that do not meet the confidence score requirements, and select the misjudged video clips to add to the video clips that meet the confidence score requirements;

[0080] According to the numbering format, the initial and end times of all video clips that meet the confidence score requirements are obtained, and all video clips are merged in chronological order to complete the online video editing.

[0081] Through the above technical solution, this embodiment first screens out several video clips based on the scene, and then, the video clips screened out in chronological order are sequentially input into the trained action recognition network model to obtain the confidence of each video clip. The editor re-screens the content of the video clips that do not meet the confidence score requirements, selects the misjudged video clips and adds them to the video clips that meet the confidence score requirements. Finally, according to the numbering format, the initial time and end time of all video clips that meet the confidence score requirements are obtained, and all video clips are merged in chronological order to complete the video online editing.

[0082] A video editing method based on content recognition, the method is applicable to the video editing system based on content recognition according to any one of claims 1 to 9, characterized in that the method comprises the following steps:

[0083] Step S1: The editor uploads the video to be edited to the online video editing client page;

[0084] Step S2: The system obtains the size and format of the video file to be edited;

[0085] Step S3, filtering out editing nodes that meet the requirements for online editing of the video to be edited according to the content obtained in step S2;

[0086] Step S4, selecting a qualified editing node according to the current state of the editing node that satisfies the online editing of the video to be edited;

[0087] Step S5, selecting the video clips that meet the requirements according to the scenes in the video content;

[0088] Step S6: Confidence scoring is performed on all the video clips that meet the requirements by using the trained model;

[0089] Step S7: select the final video clips according to the scoring results, splice them in chronological order, and complete the video editing.

[0090] All parameter calculations in the present invention are simple numerical calculations, and there is no need to consider the dimension.

[0091] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A video editing system based on content recognition, characterized in that: The system comprises: a clipping task acceptance module, a clipping task allocation module, a video clipping module and a confidence evaluation module; The editing task accepting module is used to import the video to be edited and obtain the basic data of the video to be edited; The editing task allocation module is used to allocate the editing nodes of the videos to be edited according to the basic data of each video to be edited acquired by the editing task acceptance module; The video editing module is used to edit the video based on the video content; The confidence evaluation module is used to evaluate the confidence of the clipped video clips and output the target video.

2. A video editing system based on content recognition according to claim 1, characterized in that: The working process of the editing task accepting module includes: The editor opens the online video editing client page in the browser, uploads the video file to be edited to the system and temporarily stores it in the media library; Get the size and content of each video file to be edited.

3. The video editing system based on content recognition according to claim 1, characterized in that: The working process of the editing task allocation module includes: The system obtains the size and format of the video file to be edited; Screen out the editing nodes that meet the requirements according to the size and format of the video file to be edited. If there are idle editing nodes among the editing nodes that meet the requirements, the editor manually selects any editing node for online video editing; If there is no idle node among the editing nodes that meet the requirements, the system will send online editing requests to all the editing nodes that meet the requirements in turn. When the first editing node that meets the requirements receives the online editing request, it will determine whether it is idle. If so, it will add the current video file to the editing list and perform online editing. If not, it will transfer the current editing task to the next editing node that meets the requirements. The next editing node that meets the requirements determines whether it is in an idle state. If so, it adds the current video file to the editing list for online editing. Otherwise, it continues to transfer the current editing task to the next editing node that meets the requirements. The editing task is passed in turn among all the editing nodes that meet the requirements until the current video file is added to the editing list by one of the editing nodes that meets the requirements.

4. The video editing system based on content recognition according to claim 3, characterized in that: If there is no idle node among the editing nodes that meet the requirements, and the online editing request at the system has not been responded to within the set time period, the current online editing request will be discarded and a new online editing request will be reissued.

5. The video editing system based on content recognition according to claim 3, characterized in that: The working process of screening out the clipping nodes that meet the requirements includes: Obtaining the operating parameter data of all trusted clipping nodes; The matching degree K of all trusted clip nodes and the current video file is calculated by the following formula: K = αA*βB; In the formula, α and β are weight coefficients, A is the format matching index, and B is the memory matching index; The matching degree K between the trusted clipping node and the current video file is compared with the set matching degree standard value K0. If the matching degree K between the trusted clipping node and the current video file is greater than the set matching degree standard value K0, it means that the trusted node meets the requirements, otherwise, the trusted node does not meet the requirements.

6. A computing power access capability evaluation system for supercomputing according to claim 5, characterized in that: The format matching index A is related to the format of the video file to be edited. If the format of the video file to be edited is within the video format range processed by the editing node, A=1; if the format of the video file to be edited is not within the video format range processed by the editing node, A=0.

7. A system for evaluating computing power access capability of supercomputers according to claim 5, characterized in that: The memory matching index B is related to the size of the video file to be edited. The size M of the video file to be edited is obtained. If the memory space available to the editing node is greater than 2M, B=1; if the memory space available to the editing node is less than or equal to 2M and greater than M, B=0.5; if the memory space available to the editing node is less than or equal to M, B=0.

8. The video editing system based on content recognition according to claim 1, characterized in that: The working process of the video editing module includes: Obtain video content, and extract n video clips based on specific scenes in the video content; Number each video clip in the following format: where i belongs to n, is the initial moment of the i-th video clip, is the end time of the i-th video clip, σ i is the sequence number of the i-th video clip.

9. The video editing system based on content recognition according to claim 8, characterized in that: The working process of the confidence assessment module includes: According to the initial and end times of the n video clips, the n videos are sequentially input into the trained action recognition network model, the confidence evaluation result of each video is output, and each video clip is renumbered in the following format: Among them, P i Give each a confidence score; The editors re-screen the video clips that do not meet the confidence score requirements, and select the misjudged video clips to add to the video clips that meet the confidence score requirements; According to the numbering format, the initial and end times of all video clips that meet the confidence score requirements are obtained, and all video clips are merged in chronological order to complete the online video editing.

10. A video editing method based on content recognition, the method being applicable to the video editing system based on content recognition according to any one of claims 1 to 9, characterized in that: The method comprises the following steps: Step S1: The editor uploads the video to be edited to the online video editing client page; Step S2: The system obtains the size and format of the video file to be edited; Step S3, filtering out editing nodes that meet the requirements for online editing of the video to be edited according to the content obtained in step S2; Step S4, selecting a qualified editing node according to the current state of the editing node that satisfies the online editing of the video to be edited; Step S5, selecting the video clips that meet the requirements according to the scenes in the video content; Step S6: Confidence scoring is performed on all the video clips that meet the requirements by using the trained model; Step S7: select the final video clips according to the scoring results, splice them in chronological order, and complete the video editing.

Images