WO2022253276A1 - 视频处理方法、设备、存储介质及计算机程序产品 - Google Patents

视频处理方法、设备、存储介质及计算机程序产品 Download PDF

Info

Publication number
WO2022253276A1
WO2022253276A1 PCT/CN2022/096646 CN2022096646W WO2022253276A1 WO 2022253276 A1 WO2022253276 A1 WO 2022253276A1 CN 2022096646 W CN2022096646 W CN 2022096646W WO 2022253276 A1 WO2022253276 A1 WO 2022253276A1
Authority
WO
WIPO (PCT)
Prior art keywords
video
initial
initial video
video segment
segment
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2022/096646
Other languages
English (en)
French (fr)
Inventor
何沃洲
李�根
吴兴龙
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Zitiao Network Technology Co Ltd
Original Assignee
Beijing Zitiao Network Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beijing Zitiao Network Technology Co Ltd filed Critical Beijing Zitiao Network Technology Co Ltd
Priority to US18/566,180 priority Critical patent/US12451161B2/en
Publication of WO2022253276A1 publication Critical patent/WO2022253276A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G11—INFORMATION STORAGE
    • G11B—INFORMATION STORAGE BASED ON RELATIVE MOVEMENT BETWEEN RECORD CARRIER AND TRANSDUCER
    • G11B27/00—Editing; Indexing; Addressing; Timing or synchronising; Monitoring; Measuring tape travel
    • G11B27/005—Reproducing at a different information rate from the information rate of recording
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/43—Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network or synchronising decoder's clock; Client middleware
    • H04N21/44—Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream or rendering scenes according to encoded video stream scene graphs
    • H04N21/4402—Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream or rendering scenes according to encoded video stream scene graphs involving reformatting operations of video signals for household redistribution, storage or real-time display
    • H04N21/440281—Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream or rendering scenes according to encoded video stream scene graphs involving reformatting operations of video signals for household redistribution, storage or real-time display by altering the temporal resolution, e.g. by frame skipping
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/60—Image enhancement or restoration using machine learning, e.g. neural networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/70—Denoising; Smoothing
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/20—Analysis of motion
    • G—PHYSICS
    • G11—INFORMATION STORAGE
    • G11B—INFORMATION STORAGE BASED ON RELATIVE MOVEMENT BETWEEN RECORD CARRIER AND TRANSDUCER
    • G11B27/00—Editing; Indexing; Addressing; Timing or synchronising; Monitoring; Measuring tape travel
    • G11B27/02—Editing, e.g. varying the order of information signals recorded on, or reproduced from, record carriers
    • G11B27/031—Electronic editing of digitised analogue information signals, e.g. audio or video signals
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/43—Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network or synchronising decoder's clock; Client middleware
    • H04N21/44—Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream or rendering scenes according to encoded video stream scene graphs
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/43—Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network or synchronising decoder's clock; Client middleware
    • H04N21/44—Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream or rendering scenes according to encoded video stream scene graphs
    • H04N21/4402—Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream or rendering scenes according to encoded video stream scene graphs involving reformatting operations of video signals for household redistribution, storage or real-time display
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/47—End-user applications
    • H04N21/472—End-user interface for requesting content, additional data or services; End-user interface for interacting with content, e.g. for content reservation or setting reminders, for requesting event notification, for manipulating displayed content
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/47—End-user applications
    • H04N21/472—End-user interface for requesting content, additional data or services; End-user interface for interacting with content, e.g. for content reservation or setting reminders, for requesting event notification, for manipulating displayed content
    • H04N21/47205—End-user interface for requesting content, additional data or services; End-user interface for interacting with content, e.g. for content reservation or setting reminders, for requesting event notification, for manipulating displayed content for manipulating displayed content, e.g. interacting with MPEG-4 objects, editing locally
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/80—Generation or processing of content or additional data by content creator independently of the distribution process; Content per se
    • H04N21/83—Generation or processing of protective or descriptive data associated with content; Content structuring
    • H04N21/845—Structuring of content, e.g. decomposing content into time segments
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/80—Generation or processing of content or additional data by content creator independently of the distribution process; Content per se
    • H04N21/83—Generation or processing of protective or descriptive data associated with content; Content structuring
    • H04N21/845—Structuring of content, e.g. decomposing content into time segments
    • H04N21/8456—Structuring of content, e.g. decomposing content into time segments by decomposing the content in the time domain, e.g. in time segments
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10016—Video; Image sequence
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20021—Dividing image into blocks, subimages or windows
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20084—Artificial neural networks [ANN]

Definitions

  • Embodiments of the present disclosure relate to the technical field of computer and network communication, and in particular, to a video processing method, device, storage medium, and computer program product.
  • Video speed change is a commonly used video editing method. During video editing, the user can select a certain segment of the video material, and manually perform a conventional speed change of a certain speed change magnification on the video clip, or a curve speed change of a specified curve shape.
  • Embodiments of the present disclosure provide a video processing method, device, storage medium, and computer program product to accurately determine the variable speed magnification of video clips, realize intelligent video variable speed, simplify user operations, and improve video performance capabilities without requiring users to have editing experience.
  • an embodiment of the present disclosure provides a video processing method, including:
  • the speed of the initial video segment is changed according to the speed change magnification to obtain a target video segment.
  • an embodiment of the present disclosure provides a model training method, including:
  • the training data includes sample video clips marked with variable speed
  • the initial model of the machine learning model is trained according to the training data to obtain the trained machine learning model.
  • an embodiment of the present disclosure provides a video processing device, including:
  • an acquisition unit configured to acquire an initial video segment
  • a processing unit configured to input the initial video clip into a machine learning model, and determine the speed change ratio of the initial video clip according to the output result of the machine learning model; wherein, the machine learning model has already marked the speed change magnification sample video clips were trained;
  • the speed changing unit is configured to change the speed of the initial video segment according to the speed change magnification to obtain a target video segment.
  • an embodiment of the present disclosure provides a model training device, including:
  • An acquisition unit configured to acquire a plurality of training data; the training data includes sample video clips marked with variable speed magnification;
  • the training unit is configured to train the initial model of the machine learning model according to the training data to obtain the trained machine learning model.
  • an embodiment of the present disclosure provides an electronic device, including: at least one processor and a memory;
  • the memory stores computer-executable instructions
  • the at least one processor executes the computer-executed instructions stored in the memory, so that the at least one processor executes the method described in the first aspect, the second aspect, and various possible designs of the first aspect and the second aspect.
  • the embodiments of the present disclosure provide a computer-readable storage medium, the computer-readable storage medium stores computer-executable instructions, and when the processor executes the computer-executable instructions, the above-mentioned first aspect and second aspect are realized. aspect and the method described in various possible designs of the first aspect and the second aspect.
  • the embodiments of the present disclosure provide a computer program product, including computer instructions.
  • the computer instructions are executed by a processor, the above first aspect, the second aspect, and various possible designs of the first aspect and the second aspect can be realized. described method.
  • the embodiments of the present disclosure provide a computer program.
  • the computer program is executed by a processor, the methods described in the first aspect, the second aspect and various possible designs of the first aspect and the second aspect are implemented.
  • the video processing method, device, storage medium and computer program product provided in this embodiment obtain the initial video segment; input the initial video segment into the machine learning model, and determine the variable speed magnification of the initial video segment according to the output result of the machine learning model; Among them, the machine learning model has been trained based on the sample video clips marked with the variable speed magnification; the initial video clip is changed according to the variable speed magnification to obtain the target video clip.
  • the embodiment of the present disclosure uses artificial intelligence to determine the variable speed magnification of the initial video segment through a machine learning model, reduces the dependence on user editing experience, simplifies user operations, improves processing efficiency, reduces labor costs, and effectively improves video performance. Reasonably increase the video information density, and increase the overall playback number of the video after it is released.
  • FIG. 1 is an example diagram of an application scenario of a video processing method provided by an embodiment of the present disclosure
  • FIG. 2 is a schematic flowchart of a video processing method provided by an embodiment of the present disclosure
  • FIG. 3 is a schematic flowchart of a video processing method provided by another embodiment of the present disclosure.
  • FIG. 4 is a schematic flowchart of a video processing method provided by another embodiment of the present disclosure.
  • FIG. 5 is a schematic flowchart of a video processing method provided by another embodiment of the present disclosure.
  • FIG. 6 is a structural block diagram of a video processing device provided by an embodiment of the present disclosure.
  • FIG. 7 is a structural block diagram of a model training device provided by an embodiment of the present disclosure.
  • FIG. 8 is a schematic diagram of a hardware structure of an electronic device provided by an embodiment of the present disclosure.
  • the user when performing video editing, the user usually needs to select a certain segment of the video material, and manually perform a conventional speed change of a certain speed change magnification on the video clip, or formulate a curved speed change in a curved shape, that is, the speed change magnification presents a curved distribution.
  • the variable speed ratio is greater than 1.0, it means fast playback, and if it is less than 1.0, it means slow playback.
  • VV Video View
  • the embodiment of the present disclosure considers the application of artificial intelligence to determine the variable speed ratio of the video material, so as to accurately determine the variable speed ratio of each video segment of the initial video, reduce the dependence on user editing experience, simplify user operations, and improve Processing efficiency, reducing labor costs, and effectively improving video performance capabilities, reasonably increasing video information density, and increasing the overall number of videos played after release.
  • a machine learning model can be trained, the input of which is an initial video clip, the speed change ratio of the initial video clip can be determined according to the output result of the machine learning model, and then the speed of the initial video clip can be changed according to the speed shift ratio.
  • the video processing method of the embodiment of the present disclosure can be applied in a video editing process, and a target video with a changed speed can be obtained after processing.
  • this video processing method can also be applied in the video playback process, that is, when playing a certain video, determine the variable speed magnification of a certain initial video segment of the currently played video through the above-mentioned process, according to the variable speed magnification for the initial video clip. clip to play.
  • the video processing method of the embodiment of the present disclosure is applicable to the application scenario shown in FIG.
  • the device 13 is used to provide the initial video to the terminal device 10, or if the terminal device 10 is equipped with a camera, the application scenario may only include the terminal device 10, and the terminal device 10 collects the initial video by itself; the terminal device 10 may include but not limited to Smartphones, tablet computers, notebook computers, personal computers, personal digital assistants, wearable devices, etc., the terminal device 10 can obtain initial video segments from the initial video, for example, segment the initial video, and then process it based on a machine learning model, According to the output result of the machine learning model, the variable speed ratio of the initial video clip is determined, and the speed of the initial video clip is changed according to the variable speed ratio to obtain the target video clip.
  • the machine learning model can be pre-configured on the terminal device 10; of course, it can also be configured on the cloud server.
  • the terminal device 10 can send the video frame of any video clip to the cloud server, and the cloud server can obtain the evaluation parameters corresponding to the video clip. sent to the terminal device 10.
  • the training process of the machine learning model can be performed on the terminal device 10 in advance, of course, it can also be performed on the cloud server in advance, and then the trained model is configured on the terminal device 10 .
  • FIG. 2 is a schematic flowchart of a video processing method provided by an embodiment of the present disclosure.
  • the method of this embodiment can be applied in a terminal device or a server, and the video processing method includes:
  • the initial video clip can be a video clip obtained from a video capture device or a video storage device, or a locally stored video clip, and the video clip can be a whole video or a whole video A certain segment in , there is no limit here.
  • the machine learning model may be a neural network model, or other machine learning models, which are not limited here.
  • the input of the machine learning model is the video file of the initial video segment, or the video frame extracted from the initial video segment;
  • the output of the machine learning model can be the variable speed ratio of the initial video segment, or the evaluation of the initial video segment Parameters, the evaluation parameters are parameters used to measure the appropriate speed change degree of the initial video segment. Based on the evaluation parameters of the initial video segment, the variable speed ratio of the initial video segment can be determined.
  • the output of the machine learning model is not limited to the above examples. It can be based on the machine The output of the learning model can be the variable speed ratio of the initial video clip.
  • machine learning models can be designed and trained according to different input and output, which will not be repeated here.
  • the speed change process can be performed on the initial video clip according to the speed shift ratio.
  • the editing of the initial video may be completed based on the variable speed ratio of the initial video segment, so as to generate a new video.
  • the target video can be obtained by replacing the initial video segment with the target video segment in the initial video.
  • the above process can also be applied in the video playing process, that is, when the initial video segment is played, the initial video segment is played according to the variable speed magnification of the initial video segment.
  • the video processing method provided in this embodiment obtains the initial video segment; inputs the initial video segment into the machine learning model, and determines the variable speed magnification of the initial video segment according to the output result of the machine learning model; wherein, the machine learning model has already been marked based on
  • the sample video segment with variable speed magnification is used for training; the initial video segment is changed according to the variable speed magnification to obtain the target video segment.
  • artificial intelligence is used to determine the variable speed magnification of the initial video segment through a machine learning model, which reduces the dependence on user editing experience, simplifies user operations, improves processing efficiency, reduces labor costs, and effectively improves video performance capabilities, which is reasonable. Maximize the information density of the video and increase the overall playback number of the video after it is released.
  • the acquisition of the initial video segment described in S201 includes:
  • some segments may be suitable for variable speed, and some segments may not be suitable for variable speed, so the initial video can be divided into smaller granular video segments, so that each segment The video segment judges whether the speed change is suitable, and determines the speed change magnification of the video segment suitable for speed change.
  • this embodiment considers that when changing the speed of a certain part of the video, this part of the video usually has associated parts, and the associated parts usually belong to the same scene. Therefore, when the initial video is divided into multiple video clips, according to the scene The initial video is segmented, and the same scene is used as a segment, and the initial video segment in the above embodiment may be a segment.
  • a segment may also be long, it is divided based on the segment according to the preset duration to obtain multiple video segments.
  • the preset duration can be 1 second or other durations.
  • any video segment of the shot segment can also be used as the initial video segment.
  • adjacent video segments may have a certain time overlap.
  • adjacent video segments may have a certain time overlap.
  • the first video clip can be 0-1 second
  • the second video clip can be 0.5-1.5 seconds
  • the third video clip can be 1-2 seconds, and so on.
  • adjacent video segments may not have time overlap, that is, the first video segment may be 0-1 second, the second video segment may be 1-2 seconds, and so on.
  • the duration of the segment is less than or equal to the preset duration, the segment may not be divided.
  • frame extraction can also be performed on the initial video clip, and the video frames in the initial video clip can be extracted as the input of the machine learning model.
  • the initial video clip can be processed according to a preset frame rate. Frame extraction is performed to obtain video frames in the initial video segment.
  • the preset frame rate may be 30 fps, that is, 30 video frames may be extracted from each second of the video segment.
  • in S202 inputting the initial video segment into the machine learning model, and determining the variable speed magnification of the initial video segment according to the output result of the machine learning model, Specifically, it may include:
  • variable speed ratio of the initial video segment can be obtained based on the video frame of the initial video segment based on a machine learning model.
  • the machine learning model can be a neural network model that evaluates the appropriate speed change degree of the initial video segment; the input of the model is the video frame of the initial video segment, and the output is the evaluation parameter of the video segment, and the evaluation parameter is used to represent The degree to which the initial video segment is suitable for speed change, for example, the evaluation parameter ranges from 0 to 1, the closer to 0, the more suitable for slow playback of the initial video segment, and the closer to 1, the more suitable for fast playback of the initial video segment.
  • the video clips suitable for slow playback usually have at least one of the following characteristics: the main body is clear, the camera or the main body has relatively fast movement, and can show more details and aesthetic feeling (such as the moment of pouring water, running fast and hitting a ball, etc.
  • video clips that are suitable for fast playback usually have relatively small changes between adjacent frames, contain less information, and the like.
  • the evaluation parameters corresponding to the appropriate degree of speed change can be obtained according to the input video frame of the initial video segment.
  • variable speed magnification of the initial video segment may be obtained according to the evaluation parameter of the initial video segment.
  • different evaluation parameters represent different degrees of suitable speed change for the video segment.
  • the value range of the evaluation parameter is 0 to 1, and the closer to 0, the more suitable for slow playback of the video segment, that is, the more suitable the evaluation parameter is.
  • the comparison relationship between the evaluation parameters and the variable speed magnification can be obtained in advance, and then the variable speed magnification of the initial video segment can be obtained according to the preset evaluation parameter and the variable speed magnification comparison relationship, and the evaluation parameters corresponding to the initial video segment.
  • the variable speed magnification is compared with the relationship, and the variable speed magnification corresponding to the evaluation parameter corresponding to the initial video segment is found, and used as the variable speed magnification of the initial video segment.
  • the machine learning model is a neural network model for obtaining the variable speed magnification, that is, the input of the model is the video frame of the initial video clip, and the output is the variable speed magnification of the initial video clip, that is, no longer Execute the conversion process between the evaluation parameter and the variable speed magnification separately. Therefore, based on the machine learning model, S202 is specifically inputting the video frame of the initial video segment into the machine learning model, and outputting the variable speed magnification of the initial video segment through the machine learning model.
  • the evaluation parameters of adjacent video segments can be smoothed , such as windowing and smoothing the evaluation parameters on the time axis dimension, where a triangular window or a rectangular window can be added to make the evaluation parameters change and transition more smoothly and naturally with the time axis, which in turn can make the subsequent acquired initial video clips and
  • the variable speed magnification of its adjacent video clips changes and transitions along the time axis also remains smooth and natural.
  • the speed change magnification of the initial video clip and the speed change magnification of other adjacent initial video clips may be smoothed
  • the process is to determine each smoothed variable speed ratio as the final variable speed ratio of each corresponding initial video segment.
  • variable speed magnification of the initial video segment it may further include:
  • Speed threshold for example, can be set to an average speech rate of 8 words/second, a peak speech rate of 10 words/second, etc., and then after obtaining the initial speech rate in the initial video segment, change the speed according to the initial speech rate and the target of the initial video segment Speech rate, determine the target speech rate when the initial video clip is played according to the variable speed magnification, that is, the speech rate after the speed change, and then compare the target speech rate with the preset speech rate threshold.
  • the target speech rate exceeds the preset speech rate threshold. Then adjust the variable speed magnification according to the preset speech rate threshold, so that the target speech rate when the initial video clip is played according to the adjusted variable speed magnification does not exceed the preset speech rate threshold, and determine the final value of the initial video clip with the adjusted variable speed magnification Variable speed magnification.
  • this method also includes the process of training the machine learning model, and its execution subject can be the execution subject of the above video processing method embodiment, or any other
  • the machine learning model is a neural network model for evaluating the appropriate speed change degree of video clips.
  • the training process is as follows:
  • some training data can be collected first, and the training data can include sample video clips marked with variable speed magnifications, and then the initial model of the machine learning model is trained according to the training data to obtain the trained machine learning model .
  • a sample video clip marked with a variable speed ratio can be obtained, the sample video clip is a video material that has been slowed down or fast forwarded, and marked with a slow motion tag or a fast motion tag, and/or, not Video materials that are slowed down or fast forwarded, and whose video clips are respectively marked with labels that identify the appropriate degree of speed change; video frames are extracted from the sample video clips, and the video frames are determined as the training data.
  • the real variable speed magnification can be directly marked, while for video clips that have not been slowed down or fast forwarded, and each video clip is respectively
  • the video material marked with a label indicating an appropriate speed change degree can be given an appropriate speed change ratio according to the appropriate speed change degree.
  • the training data may also be a sequence of continuous video frames labeled with preset tags, wherein, optionally, the preset tags may include tags identifying the appropriate degree of shifting or tags identifying the status of shifting, and tags identifying the status of shifting It can be the label that has been slowed down, the label that has been fastened, and the corresponding variable speed ratio can be marked; the label that identifies the appropriate speed change can be assigned a value to the continuous video frame sequence that is suitable for slow playback according to the appropriate degree , the continuous video frame sequence suitable for fast playback is assigned a value according to the degree of suitability, as the label of the video segment, as an example, four values of 0, 1, 2, and 3 can be used as the label, and the larger the value, the more suitable for slow playback , the smaller the value, the more suitable for fast playback.
  • the preset tags may include tags identifying the appropriate degree of shifting or tags identifying the status of shifting, and tags identifying the status of shifting It can be the label that has been slowed down, the label that has
  • the continuous video frame sequence is extracted from the video material, specifically, the marked video material can be obtained, and the marked video material can include: slowed down or fast forwarded, and The video material marked with a slow-motion label or a fast-motion label, and/or, a video material that is not slowed down or fast forwarded, and whose video segments are respectively marked with an appropriate speed change degree label, further, the A video frame is extracted from a video segment of the video material, and the video frame is determined as the training data.
  • the extracted video frames may carry tags marked with video material.
  • Model training is performed on video frames obtained from video materials that have been slowed down or fast forwarded and marked with slow or fast forwarding tags, so that the model can better learn real slow or fast forwarded videos What kind of characteristics does the segment have; model training can be performed on video frames obtained from video material that has not been slowed down or fast forwarded, and each video segment is marked with an appropriate speed change label, so that the model can better learn what Such video clips are suitable for speed change, and the corresponding suitable speed change degree.
  • the video material labeling process can be carried out manually, according to predetermined labeling rules, for example, for the video material, the main body is clear, the lens or the main body has a relatively fast movement, and more details and aesthetics can be displayed (such as pouring water, running fast, etc.) and the moment of hitting the ball, etc.) can be determined as video segments suitable for slow playback, while video segments with relatively small changes between adjacent frames and less information content can be determined as video segments suitable for fast playback , the video clips suitable for slow playback can be assigned a value according to the appropriate degree, and the video clip suitable for fast playback can be assigned a value according to the appropriate degree as the label of the video clip.
  • four values of 0, 1, 2, and 3 can be used as label, where the larger the value, the more suitable for slow playback, and the smaller the value, the more suitable for fast playback.
  • the initial model of the machine learning model can be trained to obtain the machine learning model.
  • the initial model of the machine learning model may be a 3D convolutional neural network, or other artificial intelligence models or machine learning models, which are not limited here, and the specific training process is not described here.
  • the marked video material can be divided into video segments, for example, the segmented segment is performed first, and the segmented segments are divided according to the preset duration, so as to obtain multiple video segments, and then each Frame extraction is performed on a video segment to obtain a video frame of each video segment.
  • the training of orderly regression is also possible, that is, first according to the continuous video frame sequence training model of the first video segment of the first video material, and then according to the continuous video of the second video segment of the first video material frame sequence training model, and then train the model according to the continuous video frame sequence of the third video segment of the first video material, and so on in sequence.
  • the probability of being greater than a certain value is obtained through the result of the ordinal regression.
  • the model can output a probability of greater than 1 as 0.8, and a probability greater than 2.
  • the probability is 0.5, and the probability greater than 3 is 0.2, and then these probabilities are weighted and transformed to obtain evaluation parameters ranging from 0 to 1.
  • variable speed magnification of each video segment of the initial video is determined by applying artificial intelligence through a machine learning model, which reduces the dependence on user editing experience, simplifies user operations, improves processing efficiency, and reduces manpower. cost, and effectively improve the video performance capability, reasonably increase the video information density, and increase the overall playback number of the video after it is released.
  • FIG. 6 is a structural block diagram of a video processing device provided in an embodiment of the present disclosure.
  • the video processing device 600 includes: an acquisition unit 601 , a processing unit 602 , and a speed change unit 603 .
  • An acquisition unit 601, configured to acquire an initial video segment
  • the processing unit 602 is configured to input the initial video clip into the machine learning model, and determine the speed change ratio of the initial video clip according to the output result of the machine learning model; wherein, the machine learning model has been marked based on The sample video clips with variable speed magnification were trained;
  • the speed change unit 603 is configured to change the speed of the initial video clip according to the speed change magnification to obtain a target video clip.
  • the obtaining unit 601 when the obtaining unit 601 obtains the initial video segment, it is configured to:
  • the initial video segment is obtained by segmenting the initial video, wherein the initial video segment is a segmented segment in the initial video.
  • the speed change unit 603 is further configured to:
  • a target video is obtained by replacing the initial video segment with the target video segment in the initial video.
  • the acquiring unit 601 when the acquiring unit 601 obtains the initial video segment by segmenting the initial video, it is configured to:
  • the shot segment is divided into a plurality of video segments, and any one of the video segments is determined as the initial video segment.
  • the processing unit 602 when the processing unit 602 inputs the initial video segment into a machine learning model, and determines the speed change ratio of the initial video segment according to an output result of the machine learning model, Used for:
  • the evaluation parameters are parameters used to measure the suitable speed change degree of the initial video segment
  • the speed change magnification of the initial video segment is determined according to the evaluation parameter corresponding to the initial video segment and the comparison relationship between the preset evaluation parameter and the speed change magnification.
  • the processing unit 602 after the processing unit 602 acquires the evaluation parameters corresponding to the initial video segment, it is further configured to:
  • variable speed magnification of the initial video segment After the determination of the variable speed magnification of the initial video segment, it also includes:
  • the processing unit 602 determines the speed change ratio of the initial video segment, it is further configured to:
  • the initial speech rate and the variable speed magnification of the initial video clip determine the target speech rate when the initial video clip is played according to the variable speed magnification
  • variable speed magnification is adjusted according to the preset speech rate threshold.
  • the video processing device provided in this embodiment can be used to execute the technical solution of the video processing method embodiment above, and its implementation principle and technical effect are similar, and will not be repeated in this embodiment.
  • FIG. 7 is a structural block diagram of a video processing device provided in an embodiment of the present disclosure.
  • the model training device 610 includes: an acquisition unit 611 and a training unit 612 .
  • An acquisition unit 611 configured to acquire a plurality of training data; the training data includes sample video clips marked with variable speed ratios;
  • the training unit 612 is configured to train an initial model of the machine learning model according to the training data to obtain a trained machine learning model.
  • the acquiring unit 611 when acquiring multiple training data, is configured to:
  • sample video segment marked with a variable speed ratio is a video material that has been slowed down or fast forwarded, and marked with a slow motion label or a fast motion label, and/or has not been slowed down or fast forwarded playback, and each video segment is marked with a label indicating the appropriate degree of speed change;
  • a video frame is extracted from the sample video segment, and the video frame is determined as the training data.
  • the model training device provided in this embodiment can be used to implement the technical solution of the above-mentioned model training method embodiment, and its implementation principle and technical effect are similar, so this embodiment will not repeat them here.
  • the electronic device 700 may be a terminal device or a server.
  • the terminal equipment may include but not limited to mobile phones, notebook computers, digital broadcast receivers, personal digital assistants (Personal Digital Assistant, PDA for short), tablet computers (Portable Android Device, PAD for short), portable multimedia players (Portable Media Player, referred to as PMP), mobile terminals such as vehicle-mounted terminals (such as vehicle-mounted navigation terminals), and fixed terminals such as digital TVs, desktop computers, etc.
  • PDA Personal Digital Assistant
  • PMP portable multimedia players
  • mobile terminals such as vehicle-mounted terminals (such as vehicle-mounted navigation terminals)
  • fixed terminals such as digital TVs, desktop computers, etc.
  • the electronic device shown in FIG. 8 is only an example, and should not limit the functions and scope of use of the embodiments of the present disclosure.
  • an electronic device 700 may include a processing device (such as a central processing unit, a graphics processing unit, etc.) 708 loads the program in the random access memory (Random Access Memory, referred to as RAM) 703 to execute various appropriate actions and processes.
  • RAM Random Access Memory
  • various programs and data necessary for the operation of the electronic device 700 are also stored.
  • the processing device 701, ROM 702, and RAM 703 are connected to each other through a bus 704.
  • An input/output (I/O) interface 705 is also connected to the bus 704 .
  • an input device 706 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; ), a speaker, a vibrator, etc.
  • a storage device 708 including, for example, a magnetic tape, a hard disk, etc.
  • the communication means 709 may allow the electronic device 700 to communicate with other devices wirelessly or by wire to exchange data. While FIG. 8 shows electronic device 700 having various means, it is to be understood that implementing or having all of the means shown is not a requirement. More or fewer means may alternatively be implemented or provided.
  • embodiments of the present disclosure include a computer program product, which includes a computer program carried on a computer-readable medium, where the computer program includes program codes for executing the methods shown in the flowcharts.
  • the computer program may be downloaded and installed from a network via communication means 709, or from storage means 708, or from ROM 702.
  • the processing device 701 the above-mentioned functions defined in the methods of the embodiments of the present disclosure are executed.
  • the above-mentioned computer-readable medium in the present disclosure may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two.
  • a computer readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to, electrical connections with one or more wires, portable computer diskettes, hard disks, random access memory (RAM), read-only memory (ROM), erasable Programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination of the above.
  • a computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.
  • a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave carrying computer-readable program code therein. Such propagated data signals may take many forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the foregoing.
  • a computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium, which can transmit, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device .
  • Program code embodied on a computer readable medium may be transmitted by any appropriate medium, including but not limited to wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.
  • the above-mentioned computer-readable medium may be included in the above-mentioned electronic device, or may exist independently without being incorporated into the electronic device.
  • the above-mentioned computer-readable medium carries one or more programs, and when the above-mentioned one or more programs are executed by the electronic device, the electronic device is made to execute the methods shown in the above-mentioned embodiments.
  • Computer program code for carrying out operations of the present disclosure can be written in one or more programming languages, or combinations thereof, including object-oriented programming languages—such as Java, Smalltalk, C++, and conventional A procedural programming language—such as "C" or a similar programming language.
  • the program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server.
  • the remote computer can be connected to the user's computer through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or it can be connected to an external A computer (connected via the Internet, eg, using an Internet service provider).
  • LAN Local Area Network
  • WAN Wide Area Network
  • each block in a flowchart or block diagram may represent a module, program segment, or portion of code that contains one or more logical functions for implementing specified executable instructions.
  • the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or they may sometimes be executed in the reverse order, depending upon the functionality involved.
  • each block of the block diagrams and/or flowchart illustrations, and combinations of blocks in the block diagrams and/or flowchart illustrations can be implemented by a dedicated hardware-based system that performs the specified functions or operations , or may be implemented by a combination of dedicated hardware and computer instructions.
  • the units involved in the embodiments described in the present disclosure may be implemented by software or by hardware. Wherein, the name of the unit does not constitute a limitation of the unit itself under certain circumstances, for example, the first obtaining unit may also be described as "a unit for obtaining at least two Internet Protocol addresses".
  • FPGAs Field Programmable Gate Arrays
  • ASICs Application Specific Integrated Circuits
  • ASSPs Application Specific Standard Products
  • SOCs System on Chips
  • CPLD Complex Programmable Logical device
  • a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device.
  • a machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium.
  • a machine-readable medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing.
  • machine-readable storage media may include one or more wire-based electrical connections, portable computer disks, hard disks, Random Access Memory (RAM), Read Only Memory (ROM), Erasable Programmable Read Only Memory (EPROM or flash memory), optical fiber, compact disk read only memory (CD-ROM), optical storage, magnetic storage, or any suitable combination of the foregoing.
  • RAM Random Access Memory
  • ROM Read Only Memory
  • EPROM Erasable Programmable Read Only Memory
  • CD-ROM compact disk read only memory
  • magnetic storage or any suitable combination of the foregoing.
  • a video processing method including:
  • the speed of the initial video segment is changed according to the speed change magnification to obtain a target video segment.
  • the acquiring the initial video segment includes:
  • the initial video segment is obtained by segmenting the initial video, wherein the initial video segment is a segmented segment in the initial video.
  • the target video segment after obtaining the target video segment, it also includes:
  • a target video is obtained by replacing the initial video segment with the target video segment in the initial video.
  • the obtaining of the initial video segment by splitting the initial video includes:
  • the shot segment is divided into a plurality of video segments, and any one of the video segments is determined as the initial video segment.
  • the inputting the initial video segment into a machine learning model, and determining the speed change ratio of the initial video segment according to an output result of the machine learning model includes:
  • the evaluation parameters are parameters used to measure the suitable speed change degree of the initial video segment
  • the speed change magnification of the initial video segment is determined according to the evaluation parameter corresponding to the initial video segment and the comparison relationship between the preset evaluation parameter and the speed change magnification.
  • after the acquisition of the evaluation parameters corresponding to the initial video segment further includes:
  • variable speed magnification of the initial video segment After the determination of the variable speed magnification of the initial video segment, it also includes:
  • variable speed ratio of the initial video segment after the determination of the variable speed ratio of the initial video segment, it further includes:
  • the initial speech rate and the variable speed magnification of the initial video clip determine the target speech rate when the initial video clip is played according to the variable speed magnification
  • variable speed magnification is adjusted according to the preset speech rate threshold.
  • a model training method comprising:
  • the training data includes sample video clips marked with variable speed
  • the initial model of the machine learning model is trained according to the training data to obtain the trained machine learning model.
  • the acquiring a plurality of training data includes:
  • sample video segment marked with a variable speed ratio is a video material that has been slowed down or fast forwarded, and marked with a slow motion label or a fast motion label, and/or has not been slowed down or fast forwarded playback, and each video segment is marked with a label indicating the appropriate degree of speed change;
  • a video frame is extracted from the sample video segment, and the video frame is determined as the training data.
  • a video processing device including:
  • an acquisition unit configured to acquire an initial video segment
  • a processing unit configured to input the initial video clip into a machine learning model, and determine the speed change ratio of the initial video clip according to the output result of the machine learning model; wherein, the machine learning model has already marked the speed change magnification sample video clips were trained;
  • the speed changing unit is configured to change the speed of the initial video segment according to the speed change magnification to obtain a target video segment.
  • the acquiring unit when the acquiring unit acquires the initial video segment, it is configured to:
  • the initial video segment is obtained by segmenting the initial video, wherein the initial video segment is a segmented segment in the initial video.
  • the speed change unit is further configured to:
  • a target video is obtained by replacing the initial video segment with the target video segment in the initial video.
  • the acquisition unit when the acquisition unit obtains the initial video segment by splitting the initial video, it is configured to:
  • the shot segment is divided into a plurality of video segments, and any one of the video segments is determined as the initial video segment.
  • the processing unit uses At:
  • the evaluation parameters are parameters used to measure the suitable speed change degree of the initial video segment
  • the speed change magnification of the initial video segment is determined according to the evaluation parameter corresponding to the initial video segment and the comparison relationship between the preset evaluation parameter and the speed change magnification.
  • the processing unit is further configured to:
  • variable speed magnification of the initial video segment After the determination of the variable speed magnification of the initial video segment, it also includes:
  • the processing unit determines the variable speed ratio of the initial video segment, it is further configured to:
  • the initial speech rate and the variable speed magnification of the initial video clip determine the target speech rate when the initial video clip is played according to the variable speed magnification
  • variable speed magnification is adjusted according to the preset speech rate threshold.
  • a model training device is provided, and the device includes:
  • An acquisition unit configured to acquire a plurality of training data; the training data includes sample video clips marked with variable speed magnification
  • the training unit is configured to train the initial model of the machine learning model according to the training data to obtain the trained machine learning model.
  • the acquiring unit when the acquiring unit acquires a plurality of training data, the acquiring unit is configured to:
  • sample video segment marked with a variable speed ratio is a video material that has been slowed down or fast forwarded, and marked with a slow motion label or a fast motion label, and/or has not been slowed down or fast forwarded playback, and each video segment is marked with a label indicating the appropriate degree of speed change;
  • a video frame is extracted from the sample video segment, and the video frame is determined as the training data.
  • an electronic device including: at least one processor and a memory;
  • the memory stores computer-executable instructions
  • the at least one processor executes the computer-executed instructions stored in the memory, so that the at least one processor executes the video processing described in the first aspect, the second aspect, and various possible designs of the first aspect and the second aspect method.
  • a computer-readable storage medium stores computer-executable instructions, and when a processor executes the computer-executable instructions, Realize the video processing method described in the first aspect, the second aspect and various possible designs of the first aspect and the second aspect.
  • a computer program product including computer instructions, when the computer instructions are executed by a processor, the above first aspect, the second aspect and the first aspect, the second aspect Aspects of various possible designs of the described video processing method.
  • a computer program is provided.
  • the processor executes the computer program, various possibilities of the above first aspect, the second aspect, and the first aspect and the second aspect can be realized.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Medical Informatics (AREA)
  • Artificial Intelligence (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Databases & Information Systems (AREA)
  • Human Computer Interaction (AREA)
  • Television Signal Processing For Recording (AREA)
  • Image Analysis (AREA)

Abstract

本公开实施例提供一种视频处理方法、设备、存储介质及计算机程序产品，通过获取初始视频片段（S201）；将初始视频片段输入机器学习模型中，并根据机器学习模型的输出结果确定初始视频片段的变速倍率；其中，机器学习模型已基于已标注了变速倍率的样本视频片段进行了训练（S202）；根据变速倍率对初始视频片段进行变速，得到目标视频片段（S203）。本公开实施例应用人工智能的方式通过机器学习模型来确定初始视频片段的变速倍率，降低对用户编辑经验的依赖，简化用户操作，提高处理效率，降低人力成本，且有效地提高视频表现能力，合理地提高视频信息密度，提高视频发布后的整体播放数。

Description

视频处理方法、设备、存储介质及计算机程序产品
相关申请交叉引用
本申请要求于2021年6月1日提交中国专利局、申请号为202110610952.5、发明名称为“视频处理方法、设备、存储介质及计算机程序产品”的中国专利申请的优先权，其全部内容通过引用并入本文。
技术领域
本公开实施例涉及计算机与网络通信技术领域，尤其涉及一种视频处理方法、设备、存储介质及计算机程序产品。
背景技术
视频变速是一种常用的视频编辑手段，在进行视频编辑时，用户可以选择视频素材的某个片段，人工对视频片段进行某一变速倍率的常规变速，或者制定曲线形状的曲线变速。
然而，通过人工方式对视频进行变速编辑往往需要用户具有一定的经验，经验不足的用户可能尝试若干次才能确定相对合适的变速效果，并且用户确定变速倍率的效率较低，最终确定的变速倍率可能达不到最佳的变速效果，使得最终得到的视频表现能力不佳。
发明内容
本公开实施例提供一种视频处理方法、设备、存储介质及计算机程序产品，以准确确定视频片段的变速倍率，实现视频智能变速，简化用户操作，不需要用户具有编辑经验，提高视频表现能力。
第一方面，本公开实施例提供一种视频处理方法，包括：
获取初始视频片段；
将所述初始视频片段输入机器学习模型中，并根据所述机器学习模型的输出结果确定所述初始视频片段的变速倍率；其中，所述机器学习模型已基于已标注了变速倍率的样本视频片段进行了训练；
根据所述变速倍率对所述初始视频片段进行变速，得到目标视频片段。
第二方面，本公开实施例提供一种模型训练方法，包括：
获取多个训练数据；所述训练数据包括已标注了变速倍率的样本视频片段；
根据所述训练数据对机器学习模型的初始模型进行训练，得到训练后的机器学习模型。
第三方面，本公开实施例提供一种视频处理设备，包括：
获取单元，用于获取初始视频片段；
处理单元，用于将所述初始视频片段输入机器学习模型中，并根据所述机器学习模型的输出结果确定所述初始视频片段的变速倍率；其中，所述机器学习模型已基于已标注了变速倍率的样本视频片段进行了训练；
变速单元，用于根据所述变速倍率对所述初始视频片段进行变速，得到目标视频片段。
第四方面，本公开实施例提供一种模型训练设备，包括：
获取单元，用于获取多个训练数据；所述训练数据包括已标注了变速倍率的样本视频片段；
训练单元，用于根据所述训练数据对机器学习模型的初始模型进行训练，得到训练后的机器学习模型。
第五方面，本公开实施例提供一种电子设备，包括：至少一个处理器和存储器；
所述存储器存储计算机执行指令；
所述至少一个处理器执行所述存储器存储的计算机执行指令，使得所述至少一个处理器执行如上第一方面、第二方面以及第一方面、第二方面各种可能的设计所述的方法。
第六方面，本公开实施例提供一种计算机可读存储介质，所述计算机可读存储介质中存储有计算机执行指令，当处理器执行所述计算机执行指令时，实现如上第一方面、第二方面以及第一方面、第二方面各种可能的设计所述的方法。
第七方面，本公开实施例提供一种计算机程序产品，包括计算机指令，该计算机指令被处理器执行时实现如上第一方面、第二方面以及第一方面、第二方面各种可能的设计所述的方法。
第八方面，本公开实施例提供一种计算机程序，该计算机程序被处理器执行时，实现如上第一方面、第二方面以及第一方面、第二方面各种可能的设计所述的方法。
本实施例提供的视频处理方法、设备、存储介质及计算机程序产品，通过获取初始视频片段；将初始视频片段输入机器学习模型中，并根据机器学习模型的输出结果确定初始视频片段的变速倍率；其中，机器学习模型已基于已标注了变速倍率的样本视频片段进行了训练；根据变速倍率对初始视频片段进行变速，得到目标视频片段。本公开实施例应用人工智能的方式通过机器学习模型来确定初始视频片段的变速倍率，降低对用户编辑经验的依赖，简化用户操作，提高处理效率，降低人力成本，且有效地提高视频表现能力，合理地提高视频信息密度，提高视频发布后的整体播放数。
附图说明
为了更清楚地说明本公开实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作一简单地介绍，显而易见地，下面描述中的附图是本公开的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1为本公开实施例提供的视频处理方法的应用场景示例图；
图2为本公开一实施例提供的视频处理方法流程示意图；
图3为本公开另一实施例提供的视频处理方法流程示意图；
图4为本公开另一实施例提供的视频处理方法流程示意图；
图5为本公开另一实施例提供的视频处理方法流程示意图；
图6为本公开一实施例提供的视频处理设备的结构框图；
图7为本公开一实施例提供的模型训练设备的结构框图；
图8为本公开一实施例提供的电子设备的硬件结构示意图。
具体实施方式
为使本公开实施例的目的、技术方案和优点更加清楚，下面将结合本公开实施例中的附图，对本公开实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例是本公开一部分实施例，而不是全部的实施例。基于本公开中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都属于本公开保护的范围。
现有技术中，在进行视频编辑时，通常需要用户选择视频素材的某个片段，人工对视频片段进行某一变速倍率的常规变速，或者制定曲线形状的曲线变速，也即变速倍率呈现曲线分布，其中通常情况下，变速倍率大于1.0为快放，小于1.0为慢放。
然而，现有技术中通过人工方式对视频进行变速编辑往往需要一定的经验，用户可能尝试若干次才能确定相对合适的变速倍率，达到相对合适的变速效果，并且用户确定变速倍率的效率较低，在面对较长的视频素材需要人工一段一段的设定变速倍率，消耗大量的人力成本和时间成本，且最终确定的变速倍率也可能达不到最佳的变速效果，使得最终得到的视频表现能力不佳，尤其是无法准确科学的提高视频的信息密度、提高视频发布后的整体播放数(Video View，简称VV，指在一个统计周期内视频被打开的总次数)。
为了解决上述技术问题，本公开实施例中考虑应用人工智能的方式来确定视频素材的变速倍率，以准确确定初始视频各视频片段的变速倍率，降低对用户编辑经验的依赖，简化用户操作，提高处理效率，降低人力成本，且有效地提高视频表现能力，合理地提高视频信息密度，提高视频发布后的整体播放数。
具体的，可训练一个机器学习模型，该模型的输入为初始视频片段，根据机器学习模型的输出结果可确定初始视频片段的变速倍率，进而可根据变速倍率对初始视频片段进行变速。
本公开实施例的视频处理方法可应用于视频编辑过程中，处理后得到经过变速的目标视频。当然，该视频处理方法也可应用在视频播放过程中，也即在播放某一视频时，通过上述过程确定当前播放的视频的某一初始视频片段的变速倍率，按照该变速倍率对该初始视频片段进行播放。
本公开实施例的视频处理方法适用于如图1所示的应用场景，包括：终端设备10，还可包括视频采集设备12和/或视频存储设备13，其中视频采集设备12和/或视频存储设备13用于向终端设备10提供初始视频，或者若终端设备10配置有摄像头，则应用场景中也可仅包括终端设备10，由终端设备10自己采集初始视频；终端设备10可包括但不限于智能手机、平板电脑、笔记本电脑、个人电脑、个人数字助理、可穿戴设备等等，终端设备10可将初始视频中得到初始视频片段，例如对初始视频进行分割，进而基于机器学习模型进行处理，根据机器学习模型的输出结果确定初始视频片段的变速倍率，根据变速倍率对初始视频片段进行变速，得到目标视频片段。
其中机器学习模型可预先配置在终端设备10上；当然也可配置在云端服务器上，终端设备10可将任一视频片段的视频帧发送给云端服务器，由云端服务器获取视频片段对应的评价参数后发送给终端设备10。此外对于机器学习模型的训练过程可以预先在终端设备10上进行，当然也可预先在云端服务器上进行，再将训练好的模型配置在终端设备10上。
下面将结合具体实施例对本公开提供的视频处理方法流程进行详细介绍。
参考图2，图2为本公开实施例提供的视频处理方法流程示意图。本实施例的方法可以应用在终端设备或服务器中，该视频处理方法包括：
S201、获取初始视频片段。
在本实施例中，初始视频片段可以为从视频采集设备或视频存储设备获取到的视频片段，也可以是本地存储的视频片段，视频片段可以是一整段视频，也可以是一整段视频中的某一个片段，此处可不做限制。
S202、将所述初始视频片段输入机器学习模型中，并根据所述机器学习模型的输出结果确定所述初始视频片段的变速倍率；其中，所述机器学习模型已基于已标注了变速倍率的样本视频片段进行了训练。
在本实施例中，机器学习模型可以为神经网络模型，或者其他机器学习模型，此处可不做限制。此外，机器学习模型的输入为初始视频片段的视频文件，也可以为从初始视频片段中提取的视频帧；机器学习模型的输出可以为初始视频片段的变速倍率，也可以为初始视频片段的评价参数，评价参数为用于衡量初始视频片段适宜变速程度的大小的参数，基于初始视频片段的评价参数可确定初始视频片段的变速倍率，当然机器学习模型的输出也可不限于上述举例，可基于机器学习模型的输出得到初始视频片段的变速倍率即可。相应的，可根据输入和输出的不同设计和训练机器学习模型，此处不再赘述。
S203、根据所述变速倍率对所述初始视频片段进行变速，得到目标视频片段。
在本实施例中，在得到初始视频片段的变速倍率后，可根据变速倍率对初始视频片段进行变速处理。
可选的，在视频编辑过程中，可基于初始视频片段的变速倍率，完成对初始视频的编辑，从而生成新视频。具体的，初始视频片段来源于初始视频时，可在得到目标视频片段之后，通过在初始视频中将初始视频片段替换为所述目标视频片段，得到目标视频。
当然，上述过程也可应用在视频播放过程中，也即在播放初始视频片段时，根据初始视频片段的变速倍率对初始视频片段进行播放。
本实施例提供的视频处理方法，通过获取初始视频片段；将初始视频片段输入机器学习模型中，并根据机器学习模型的输出结果确定初始视频片段的变速倍率；其中，机器学习模型已基于已标注了变速倍率的样本视频片段进行了训练；根据变速倍率对初始视频片段进行变速，得到目标视频片段。本实施例应用人工智能的方式通过机器学习模型来确定初始视频片段的变速倍率，降低对用户编辑经验的依赖，简化用户操作，提高处理效率，降低人力成本，且有效地提高视频表现能力，合理地提高视频信息密度，提高视频发布后的整体播放数。
在上述实施例的基础上，S201所述的获取初始视频片段，包括：
S2021、获取初始视频；
S2022、通过对所述初始视频进行分镜，得到所述初始视频片段，其中，所述初始视频片段为所述初始视频中的分镜片段。
在本实施例中，对于初始视频，可能某些片段适宜变速，某些片段不适宜变速，因此可对初始视频进行分割，将初始视频分割成更小粒度的视频片段，以便于分别对每一段视频片段判断是否适宜变速，并确定适宜变速的视频片段的变速倍率。
可选的，本实施例考虑到对视频某一部分进行变速时该部分视频通常具有关联的部分，关联的部分通常属于同一个场景，因此在将初始视频分割为多个视频片段时，可根据场景对所述初始视频进行分镜，同一场景作为一个分镜片段，上述实施例中的初始视频片段可以为某一分镜片段。
可选的，考虑到一个分镜片段也可能较长，因此在分镜片段的基础上按照预设时长进行分割，得到多个视频片段，例如预设时长可以为1秒或者其他时长，本实施例中还可将分镜片段的任一视频片段作为初始视频片段。
可选的，在对初始视频的分镜片段整段进行编辑的过程中，需要获取分镜片段的每一视频片段的变速倍率。可选的，在对分镜片段分割时，相邻的视频片段可具有一定的时间交叠，以预设时长为1秒为例，本实施例中相邻的视频片段可具有一定的时间交叠，例如第一段视频片段可以为0～1秒，第二段视频片段可以为0.5～1.5秒，第三段视频片段可以为1～2秒，以此类推，通过时间交叠，可以在后续过程中更加准确地确定各视频片段适宜变速的程度，但处理量相对增加，仅此时间交叠的大小可根据处理量和效果等多种因素进行设定。当然，相邻的视频片段也可不具有时间交叠，也即第一段视频片段可以为0～1秒，第二段视频片段可以为1～2秒，以此类推。此外，若分镜片段的时长小于或等于预设时长，则可不对分镜片段进行分割。
进一步的，在上述实施例的基础上，还可对初始视频片段进行抽帧，提取初始视频片段中的视频帧作为机器学习模型的输入，可选的，可以按照预设帧率对初始视频片段进行抽帧，得到初始视频片段中的视频帧，例如预设帧率可以为30fps，也即每一秒视频片段中可以抽取到30个视频帧。
在上述实施例的基础上，如图3所示，S202所述的将所述初始视频片段输入机器学习模型中，并根据所述机器学习模型的输出结果确定所述初始视频片段的变速倍率，具体可包括：
S301、提取所述初始视频片段的视频帧；
S302、将所述初始视频片段的视频帧输入机器学习模型中，获取所述初始视频片段对应的评价参数，所述评价参数为用于衡量所述初始视频片段适宜变速程度的大小的参数；
S303、根据所述初始视频片段对应的评价参数、以及预设评价参数与变速倍率对照关系，确定所述初始视频片段的变速倍率。
在本实施例中，可基于机器学习模型来实现根据所述初始视频片段的视频帧，得到初始视频片段的变速倍率。
可选的，该机器学习模型可以为评价初始视频片段适宜变速程度的神经网络模型；该模型的输入为初始视频片段的视频帧，输出为对该视频片段的评价参数，该评价参数用于表示初始视频片段适宜变速的程度，例如评价参数的取值范围为0～1，越接近0表示对该初始视频片段越适宜慢放，越接近1表示对该初始视频片段越适宜快放。其中，适宜慢放的视频片段通常具有如下至少一种特征：主体清晰、镜头或主体有较快速的运动、能展示出更多的细节和美感(例如倒水、快跑和击球的瞬间等等)；而适宜快放的视频片段通常相邻帧之间变化比较小、包含信息量比较少等等。通过训练神经网络模型进行学习，可以根据输入的初始视频片段的视频帧得到对应的适宜变速程度的评价参数。
进一步的，在获取到初始视频片段的评价参数后，可根据该初始视频片段的评价参数获取初始视频片段的变速倍率。在上述实施例中，评价参数不同代表视频片段适宜变速的程度不同，尤其是，评价参数的取值范围为0～1，越接近0表示对该视频片段越适宜慢放，也即评价参数越小对该视频片段越适宜慢放，越接近1表示对该视频片段越适宜快放，也即评价参数越大对该视频片段越适宜快放，评价参数与变速倍率存在对照关系，本实施例中可预先获取评价参数与变速倍率对照关系，进而可根据预设的评价参数与变速倍率对照关系、以及初 始视频片段对应的评价参数，获取初始视频片段的变速倍率，也即通过查评价参数与变速倍率对照关系，找到初始视频片段对应的评价参数对应的变速倍率，作为所述初始视频片段的变速倍率。
在另一种可选实施例中，该机器学习模型为获取变速倍率的神经网络模型，也即该模型的输入为初始视频片段的视频帧，输出为初始视频片段的变速倍率，也即不再单独执行评价参数到变速倍率之间的转换过程。因此基于该机器学习模型，S202具体为将所述初始视频片段的视频帧输入机器学习模型，通过所述机器学习模型输出所述初始视频片段的变速倍率。
在上述任一实施例的基础上，可选的，在S302所述获取所述初始视频片段对应的评价参数后，还可包括：
对所述初始视频片段对应的评价参数以及其相邻的其他初始视频片段对应的评价参数进行平滑处理，将平滑处理后的各评价参数确定为对应的各初始视频片段最终的评价参数。
在本实施例中，为了避免初始视频片段及其相邻视频片段在进行变速时存在突然变速、导致过渡不平滑不自然，影响视频表现能力，可以对相邻的视频片段的评价参数进行平滑处理，例如在时间轴维度上对评价参数进行加窗平滑处理，其中可以加三角窗或者矩形窗，使得评价参数随时间轴变化和过渡更加平滑和自然，进而可以使得后续获取到的初始视频片段及其相邻视频片段的变速倍率随时间轴变化和过渡也保持平滑和自然。
在另一种可选实施例中，可以在S303所述的确定所述初始视频片段的变速倍率后，对所述初始视频片段的变速倍率以及其相邻的其他初始视频片段的变速倍率进行平滑处理，将平滑处理后的各变速倍率确定为对应的各初始视频片段最终的变速倍率。通过本实施例对变速倍率进行平滑处理，也可以使得初始视频片段及其相邻视频片段的变速倍率随时间轴变化和过渡也保持平滑和自然。其中，平滑处理过程可同上述实施例，此处不再赘述。
在上述任一实施例的基础上，如图4所示，在确定所述初始视频片段的变速倍率后，还可包括：
S401、获取所述初始视频片段中的初始语速；
S402、根据所述初始语速以及所述初始视频片段的变速倍率，确定所述初始视频片段按照所述变速倍率播放时的目标语速；
S403、若所述目标语速超过预设语速阈值，则根据所述预设语速阈值调整所述变速倍率。
在本实施例中，由于一些初始视频片段可能伴随有语音，为了避免初始视频片段变速后语速过快导致信息量过分密集，或者语速过慢导致信息量过分稀疏，因此可以配置预设语速阈值，例如可以设置为平均语速8字/秒、峰值语速10字/秒等，进而在获取到初始视频片段中的初始语速后，根据初始语速以及该初始视频片段的目标变速语速，确定初始视频片段按照变速倍率播放时的目标语速，也即变速之后的语速，然后将目标语速与预设语速阈值进行比较，若目标语速超过预设语速阈值，则根据预设语速阈值调整变速倍率，使得初始视频片段按照调整后的变速倍率播放时的目标语速不超过预设语速阈值，并以调整后的变速倍率确定为初始视频片段的最终的变速倍率。
在上述任一实施例的基础上，如图5所示，本方法还包括对机器学习模型进行训练的过程，其执行主体可以为上述视频处理方法实施例的执行主体，也可以为其他任一终端设备或服务器等电子设备，其中机器学习模型为评价视频片段适宜变速程度的神经网络模型，训练过程具体如下：
S501、获取多个训练数据；所述训练数据包括已标注了变速倍率的样本视频片段；
S502、根据所述训练数据对机器学习模型的初始模型进行训练，得到训练后的机器学习模型。
在本实施例中，可以首先采集一些训练数据，所述训练数据可以包括已标注了变速倍率的样本视频片段，进而根据训练数据对机器学习模型的初始模型进行训练，得到训练后的机器学习模型。
可选的，可获取已标注了变速倍率的样本视频片段，所述样本视频片段为已被慢放或快放、且被标注有慢放标签或快放标签的视频素材，和/或，未被慢放或快放、且其各视频片段分别被标注有标识适宜变速程度的标签的视频素材；对所述样本视频片段提取视频帧，将所述视频帧确定为所述训练数据。其中，对于已被慢放或快放、且被标注有慢放标签或快放标签的视频素材，可以直接标注真实的变速倍率，而对于未被慢放或快放、且其各视频片段分别被标注有标识适宜变速程度的标签的视频素材，可以根据适宜变速程度赋予其合适的变速倍率。
可选的，训练数据还可以为标注预设标签的连续视频帧序列，其中，可选的，所述预设标签可包括标识适宜变速程度的标签或者标识变速情况的标签，标识变速情况的标签可以为标识已被慢放的标签、标识已被快放的标签，进一步还可标注对应的变速倍率；而标识适宜变速程度的标签可以是对适宜慢放的连续视频帧序列按照适宜程度赋予数值，对适宜快放的连续视频帧序列按照适宜程度赋予数值，作为视频片段的标签，作为示例，可以以0、1、2、3四个数值作为标签，其中，数值越大表示越适宜慢放，数值越小表示越适宜快放。
在一种可选实施例中，连续视频帧序列是通过从视频素材中提取的，具体的，可获取经标注的视频素材，经过标注的视频素材可以包括：已被慢放或快放、且被标注有慢放标签或快放标签的视频素材，和/或，未被慢放或快放、且其各视频片段分别被标注有适宜变速程度标签的视频素材，进一步的，可对所述视频素材的视频片段提取视频帧，将所述视频帧确定为所述训练数据。其中提取的视频帧可以携带视频素材标注的标签。通过已被慢放或快放、且被标注有慢放标签或快放标签的视频素材得到的视频帧进行模型训练，可以使得模型能够更好的学习到真实的被慢放或快放的视频片段具有怎样的特征；而通过未被慢放或快放、且其各视频片段分别被标注有适宜变速程度标签的视频素材得到的视频帧进行模型训练，可以使得模型能够更好的学习到什么样的视频片段适宜变速、以及对应的适宜变速程度。
其中视频素材标注过程可以人为进行，根据预定的标注规则进行标注，例如对于视频素材中主体清晰、镜头或主体有较快速的运动、能展示出更多的细节和美感(例如倒水、快跑和击球的瞬间等等)的视频片段，可以确定为适宜慢放的视频片段，而相邻帧之间变化比较小、包含信息量比较少的视频片段，可以确定为适宜快放的视频片段，可对适宜慢放的视频片段按照适宜程度赋予数值，对适宜快放的视频片段按照适宜程度赋予数值，作为视频片段的标签，作为示例，可以以0、1、2、3四个数值作为标签，其中，数值越大表示越适宜慢放，数值越小表示越适宜快放。
在获取到训练数据后，可对机器学习模型的初始模型进行训练，得到所述机器学习模型。其中机器学习模型的初始模型可以为3D卷积神经网络，或者也可以为其他人工智能模型、机器学习模型，此处不做限定，具体的训练过程此处也不做赘述。需要说明的是，本实施例中对于经过标注的视频素材可以进行视频片段的分割，例如先进行分镜，并将分镜片段按照预 设时长进行分割，从而得到多个视频片段，进而对每一视频片段进行抽帧，得到每一视频片段的视频帧。
本实施例中在训练机器学习模型的初始模型时，不同视频素材的视频片段的连续视频帧序列之间不限顺序，可以随机打乱，例如可以先根据第一视频素材的第二视频片段的连续视频帧序列训练模型，再根据第二视频素材的第一视频片段的连续视频帧序列训练模型，再根据第一视频素材的第一视频片段的连续视频帧序列训练模型等等，不限制训练顺序；当然本实施例中也可有序回归的训练，也即先根据第一视频素材的第一视频片段的连续视频帧序列训练模型，再根据第一视频素材的第二视频片段的连续视频帧序列训练模型，然后再根据第一视频素材的第三视频片段的连续视频帧序列训练模型，以此类推按顺序训练。在上述示例中，由于视频片段的标签为离散的数值，通过有序回归的结果时得到大于某数值的概率，例如对于某一视频片段，模型可以输出其大于1的概率是0.8，大于2的概率是0.5，大于3的概率是0.2，然后将这些概率通过加权变换后得到0～1范围的评价参数。
通过上述各实施例提供的视频处理方法，应用人工智能的方式通过机器学习模型来确定初始视频的各视频片段的变速倍率，降低对用户编辑经验的依赖，简化用户操作，提高处理效率，降低人力成本，且有效地提高视频表现能力，合理地提高视频信息密度，提高视频发布后的整体播放数。
对应于上文实施例的视频处理方法，图6为本公开实施例提供的视频处理设备的结构框图。为了便于说明，仅示出了与本公开实施例相关的部分。参照图6，所述视频处理设备600包括：获取单元601、处理单元602、变速单元603。
获取单元601，用于获取初始视频片段；
处理单元602，用于将所述初始视频片段输入机器学习模型中，并根据所述机器学习模型的输出结果确定所述初始视频片段的变速倍率；其中，所述机器学习模型已基于已标注了变速倍率的样本视频片段进行了训练；
变速单元603，用于根据所述变速倍率对所述初始视频片段进行变速，得到目标视频片段。
根据本公开的一个或多个实施例，所述获取单元601在获取初始视频片段时，用于：
获取初始视频；
通过对所述初始视频进行分镜，得到所述初始视频片段，其中，所述初始视频片段为所述初始视频中的分镜片段。
根据本公开的一个或多个实施例，所述变速单元603在得到目标视频片段之后，还用于：
通过在所述初始视频中将所述初始视频片段替换为所述目标视频片段，得到目标视频。
根据本公开的一个或多个实施例，所述获取单元601在通过对所述初始视频进行分镜，得到所述初始视频片段时，用于：
对所述初始视频进行分镜，得到所述初始视频中的分镜片段；
将所述分镜片段分割为多个视频片段，将其中任一视频片段确定为所述初始视频片段。
根据本公开的一个或多个实施例，所述处理单元602在将所述初始视频片段输入机器学习模型中，并根据所述机器学习模型的输出结果确定所述初始视频片段的变速倍率时，用于：
提取所述初始视频片段的视频帧；
将所述初始视频片段的视频帧输入机器学习模型中，获取所述初始视频片段对应的评价参数，所述评价参数为用于衡量所述初始视频片段适宜变速程度的大小的参数；
根据所述初始视频片段对应的评价参数、以及预设评价参数与变速倍率对照关系，确定所述初始视频片段的变速倍率。
根据本公开的一个或多个实施例，所述处理单元602在获取所述初始视频片段对应的评价参数后，还用于：
对所述初始视频片段对应的评价参数以及其相邻的其他初始视频片段对应的评价参数进行平滑处理，将平滑处理后的各评价参数确定为对应的各初始视频片段最终的评价参数；或者
在所述确定所述初始视频片段的变速倍率后，还包括：
对所述初始视频片段的变速倍率以及其相邻的其他初始视频片段的变速倍率进行平滑处理，将平滑处理后的各变速倍率确定为对应的各初始视频片段最终的变速倍率。
根据本公开的一个或多个实施例，所述处理单元602在确定所述初始视频片段的变速倍率后，还用于：
获取所述初始视频片段中的初始语速；
根据所述初始语速以及所述初始视频片段的变速倍率，确定所述初始视频片段按照所述变速倍率播放时的目标语速；
若所述目标语速超过预设语速阈值，则根据所述预设语速阈值调整所述变速倍率。
本实施例提供的视频处理设备，可用于执行上述视频处理方法实施例的技术方案，其实现原理和技术效果类似，本实施例此处不再赘述。
对应于上文实施例的视频处理方法，图7为本公开实施例提供的视频处理设备的结构框图。为了便于说明，仅示出了与本公开实施例相关的部分。参照图7，模型训练设备610所述设备包括：获取单元611以及训练单元612。
获取单元611，用于获取多个训练数据；所述训练数据包括已标注了变速倍率的样本视频片段；
训练单元612，用于根据所述训练数据对机器学习模型的初始模型进行训练，得到训练后的机器学习模型。
根据本公开的一个或多个实施例，所述获取单元611在获取多个训练数据时，用于：
获取已标注了变速倍率的样本视频片段，所述样本视频片段为已被慢放或快放、且被标注有慢放标签或快放标签的视频素材，和/或，未被慢放或快放、且其各视频片段分别被标注有标识适宜变速程度的标签的视频素材；
对所述样本视频片段提取视频帧，将所述视频帧确定为所述训练数据。本实施例提供的模型训练设备，可用于执行上述模型训练方法实施例的技术方案，其实现原理和技术效果类似，本实施例此处不再赘述。
参考图8，其示出了适于用来实现本公开实施例的电子设备700的结构示意图，该电子设备700可以为终端设备或服务器。其中，终端设备可以包括但不限于诸如移动电话、笔记本电脑、数字广播接收器、个人数字助理(Personal Digital Assistant，简称PDA)、平板电脑(Portable Android Device，简称PAD)、便携式多媒体播放器(Portable Media Player，简称PMP)、车 载终端(例如车载导航终端)等等的移动终端以及诸如数字TV、台式计算机等等的固定终端。图8示出的电子设备仅仅是一个示例，不应对本公开实施例的功能和使用范围带来任何限制。
如图8所示，电子设备700可以包括处理装置(例如中央处理器、图形处理器等)701，其可以根据存储在只读存储器(Read Only Memory，简称ROM)702中的程序或者从存储装置708加载到随机访问存储器(Random Access Memory，简称RAM)703中的程序而执行各种适当的动作和处理。在RAM 703中，还存储有电子设备700操作所需的各种程序和数据。处理装置701、ROM 702以及RAM 703通过总线704彼此相连。输入/输出(I/O)接口705也连接至总线704。
通常，以下装置可以连接至I/O接口705：包括例如触摸屏、触摸板、键盘、鼠标、摄像头、麦克风、加速度计、陀螺仪等的输入装置706；包括例如液晶显示器(Liquid Crystal Display，简称LCD)、扬声器、振动器等的输出装置707；包括例如磁带、硬盘等的存储装置708；以及通信装置709。通信装置709可以允许电子设备700与其他设备进行无线或有线通信以交换数据。虽然图8示出了具有各种装置的电子设备700，但是应理解的是，并不要求实施或具备所有示出的装置。可以替代地实施或具备更多或更少的装置。
特别地，根据本公开的实施例，上文参考流程图描述的过程可以被实现为计算机软件程序。例如，本公开的实施例包括一种计算机程序产品，其包括承载在计算机可读介质上的计算机程序，该计算机程序包含用于执行流程图所示的方法的程序代码。在这样的实施例中，该计算机程序可以通过通信装置709从网络上被下载和安装，或者从存储装置708被安装，或者从ROM 702被安装。在该计算机程序被处理装置701执行时，执行本公开实施例的方法中限定的上述功能。
需要说明的是，本公开上述的计算机可读介质可以是计算机可读信号介质或者计算机可读存储介质或者是上述两者的任意组合。计算机可读存储介质例如可以是——但不限于——电、磁、光、电磁、红外线、或半导体的系统、装置或器件，或者任意以上的组合。计算机可读存储介质的更具体的例子可以包括但不限于：具有一个或多个导线的电连接、便携式计算机磁盘、硬盘、随机访问存储器(RAM)、只读存储器(ROM)、可擦式可编程只读存储器(EPROM或闪存)、光纤、便携式紧凑磁盘只读存储器(CD-ROM)、光存储器件、磁存储器件、或者上述的任意合适的组合。在本公开中，计算机可读存储介质可以是任何包含或存储程序的有形介质，该程序可以被指令执行系统、装置或者器件使用或者与其结合使用。而在本公开中，计算机可读信号介质可以包括在基带中或者作为载波一部分传播的数据信号，其中承载了计算机可读的程序代码。这种传播的数据信号可以采用多种形式，包括但不限于电磁信号、光信号或上述的任意合适的组合。计算机可读信号介质还可以是计算机可读存储介质以外的任何计算机可读介质，该计算机可读信号介质可以发送、传播或者传输用于由指令执行系统、装置或者器件使用或者与其结合使用的程序。计算机可读介质上包含的程序代码可以用任何适当的介质传输，包括但不限于：电线、光缆、RF(射频)等等，或者上述的任意合适的组合。
上述计算机可读介质可以是上述电子设备中所包含的；也可以是单独存在，而未装配入该电子设备中的。
上述计算机可读介质承载有一个或者多个程序，当上述一个或者多个程序被该电子设备执行时，使得该电子设备执行上述实施例所示的方法。
可以以一种或多种程序设计语言或其组合来编写用于执行本公开的操作的计算机程序代码，上述程序设计语言包括面向对象的程序设计语言——诸如Java、Smalltalk、C++，还包括常规的过程式程序设计语言——诸如“C”语言或类似的程序设计语言。程序代码可以完全地在用户计算机上执行、部分地在用户计算机上执行、作为一个独立的软件包执行、部分在用户计算机上部分在远程计算机上执行、或者完全在远程计算机或服务器上执行。在涉及远程计算机的情形中，远程计算机可以通过任意种类的网络——包括局域网(Local Area Network，简称LAN)或广域网(Wide Area Network，简称WAN)—连接到用户计算机，或者，可以连接到外部计算机(例如利用因特网服务提供商来通过因特网连接)。
附图中的流程图和框图，图示了按照本公开各种实施例的系统、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上，流程图或框图中的每个方框可以代表一个模块、程序段、或代码的一部分，该模块、程序段、或代码的一部分包含一个或多个用于实现规定的逻辑功能的可执行指令。也应当注意，在有些作为替换的实现中，方框中所标注的功能也可以以不同于附图中所标注的顺序发生。例如，两个接连地表示的方框实际上可以基本并行地执行，它们有时也可以按相反的顺序执行，这依所涉及的功能而定。也要注意的是，框图和/或流程图中的每个方框、以及框图和/或流程图中的方框的组合，可以用执行规定的功能或操作的专用的基于硬件的系统来实现，或者可以用专用硬件与计算机指令的组合来实现。
描述于本公开实施例中所涉及到的单元可以通过软件的方式实现，也可以通过硬件的方式来实现。其中，单元的名称在某种情况下并不构成对该单元本身的限定，例如，第一获取单元还可以被描述为“获取至少两个网际协议地址的单元”。
本文中以上描述的功能可以至少部分地由一个或多个硬件逻辑部件来执行。例如，非限制性地，可以使用的示范类型的硬件逻辑部件包括：现场可编程门阵列(FPGA)、专用集成电路(ASIC)、专用标准产品(ASSP)、片上系统(SOC)、复杂可编程逻辑设备(CPLD)等等。
在本公开的上下文中，机器可读介质可以是有形的介质，其可以包含或存储以供指令执行系统、装置或设备使用或与指令执行系统、装置或设备结合地使用的程序。机器可读介质可以是机器可读信号介质或机器可读储存介质。机器可读介质可以包括但不限于电子的、磁性的、光学的、电磁的、红外的、或半导体系统、装置或设备，或者上述内容的任何合适组合。机器可读存储介质的更具体示例可以包括基于一个或多个线的电气连接、便携式计算机盘、硬盘、随机存取存储器(RAM)、只读存储器(ROM)、可擦除可编程只读存储器(EPROM或快闪存储器)、光纤、便捷式紧凑盘只读存储器(CD-ROM)、光学储存设备、磁储存设备、或上述内容的任何合适组合。
第一方面，根据本公开的一个或多个实施例，提供了一种视频处理方法，包括：
获取初始视频片段；
将所述初始视频片段输入机器学习模型中，并根据所述机器学习模型的输出结果确定所述初始视频片段的变速倍率；其中，所述机器学习模型已基于已标注了变速倍率的样本视频片段进行了训练；
根据所述变速倍率对所述初始视频片段进行变速，得到目标视频片段。
根据本公开的一个或多个实施例，所述获取初始视频片段，包括：
获取初始视频；
通过对所述初始视频进行分镜，得到所述初始视频片段，其中，所述初始视频片段为所述初始视频中的分镜片段。
根据本公开的一个或多个实施例，在得到目标视频片段之后，还包括：
通过在所述初始视频中将所述初始视频片段替换为所述目标视频片段，得到目标视频。
根据本公开的一个或多个实施例，所述通过对所述初始视频进行分镜，得到所述初始视频片段，包括：
对所述初始视频进行分镜，得到所述初始视频中的分镜片段；
将所述分镜片段分割为多个视频片段，将其中任一视频片段确定为所述初始视频片段。
根据本公开的一个或多个实施例，所述将所述初始视频片段输入机器学习模型中，并根据所述机器学习模型的输出结果确定所述初始视频片段的变速倍率，包括：
提取所述初始视频片段的视频帧；
将所述初始视频片段的视频帧输入机器学习模型中，获取所述初始视频片段对应的评价参数，所述评价参数为用于衡量所述初始视频片段适宜变速程度的大小的参数；
根据所述初始视频片段对应的评价参数、以及预设评价参数与变速倍率对照关系，确定所述初始视频片段的变速倍率。
根据本公开的一个或多个实施例，在所述获取所述初始视频片段对应的评价参数后，还包括：
对所述初始视频片段对应的评价参数以及其相邻的其他初始视频片段对应的评价参数进行平滑处理，将平滑处理后的各评价参数确定为对应的各初始视频片段最终的评价参数；或者
在所述确定所述初始视频片段的变速倍率后，还包括：
对所述初始视频片段的变速倍率以及其相邻的其他初始视频片段的变速倍率进行平滑处理，将平滑处理后的各变速倍率确定为对应的各初始视频片段最终的变速倍率。
根据本公开的一个或多个实施例，在所述确定所述初始视频片段的变速倍率后，还包括：
获取所述初始视频片段中的初始语速；
根据所述初始语速以及所述初始视频片段的变速倍率，确定所述初始视频片段按照所述变速倍率播放时的目标语速；
若所述目标语速超过预设语速阈值，则根据所述预设语速阈值调整所述变速倍率。
第二方面，根据本公开的一个或多个实施例，提供了一种模型训练方法，所述方法包括：
获取多个训练数据；所述训练数据包括已标注了变速倍率的样本视频片段；
根据所述训练数据对机器学习模型的初始模型进行训练，得到训练后的机器学习模型。
根据本公开的一个或多个实施例，所述获取多个训练数据，包括：
获取已标注了变速倍率的样本视频片段，所述样本视频片段为已被慢放或快放、且被标注有慢放标签或快放标签的视频素材，和/或，未被慢放或快放、且其各视频片段分别被标注有标识适宜变速程度的标签的视频素材；
对所述样本视频片段提取视频帧，将所述视频帧确定为所述训练数据。
第三方面，根据本公开的一个或多个实施例，提供了一种视频处理设备，包括：
获取单元，用于获取初始视频片段；
处理单元，用于将所述初始视频片段输入机器学习模型中，并根据所述机器学习模型的输出结果确定所述初始视频片段的变速倍率；其中，所述机器学习模型已基于已标注了变速倍率的样本视频片段进行了训练；
变速单元，用于根据所述变速倍率对所述初始视频片段进行变速，得到目标视频片段。
根据本公开的一个或多个实施例，所述获取单元在获取初始视频片段时，用于：
获取初始视频；
通过对所述初始视频进行分镜，得到所述初始视频片段，其中，所述初始视频片段为所述初始视频中的分镜片段。
根据本公开的一个或多个实施例，所述变速单元在得到目标视频片段之后，还用于：
通过在所述初始视频中将所述初始视频片段替换为所述目标视频片段，得到目标视频。
根据本公开的一个或多个实施例，所述获取单元在通过对所述初始视频进行分镜，得到所述初始视频片段时，用于：
对所述初始视频进行分镜，得到所述初始视频中的分镜片段；
将所述分镜片段分割为多个视频片段，将其中任一视频片段确定为所述初始视频片段。
根据本公开的一个或多个实施例，所述处理单元在将所述初始视频片段输入机器学习模型中，并根据所述机器学习模型的输出结果确定所述初始视频片段的变速倍率时，用于：
提取所述初始视频片段的视频帧；
将所述初始视频片段的视频帧输入机器学习模型中，获取所述初始视频片段对应的评价参数，所述评价参数为用于衡量所述初始视频片段适宜变速程度的大小的参数；
根据所述初始视频片段对应的评价参数、以及预设评价参数与变速倍率对照关系，确定所述初始视频片段的变速倍率。
根据本公开的一个或多个实施例，所述处理单元在获取所述初始视频片段对应的评价参数后，还用于：
对所述初始视频片段对应的评价参数以及其相邻的其他初始视频片段对应的评价参数进行平滑处理，将平滑处理后的各评价参数确定为对应的各初始视频片段最终的评价参数；或者
在所述确定所述初始视频片段的变速倍率后，还包括：
对所述初始视频片段的变速倍率以及其相邻的其他初始视频片段的变速倍率进行平滑处理，将平滑处理后的各变速倍率确定为对应的各初始视频片段最终的变速倍率。
根据本公开的一个或多个实施例，所述处理单元在确定所述初始视频片段的变速倍率后，还用于：
获取所述初始视频片段中的初始语速；
根据所述初始语速以及所述初始视频片段的变速倍率，确定所述初始视频片段按照所述变速倍率播放时的目标语速；
若所述目标语速超过预设语速阈值，则根据所述预设语速阈值调整所述变速倍率。
第四方面，根据本公开的一个或多个实施例，提供了一种模型训练设备，所述设备包括：
获取单元，用于获取多个训练数据；所述训练数据包括已标注了变速倍率的样本视频片段
训练单元，用于根据所述训练数据对机器学习模型的初始模型进行训练，得到训练后的机器学习模型。
根据本公开的一个或多个实施例，所述获取单元在获取单元在获取多个训练数据时，用于：
获取已标注了变速倍率的样本视频片段，所述样本视频片段为已被慢放或快放、且被标注有慢放标签或快放标签的视频素材，和/或，未被慢放或快放、且其各视频片段分别被标注有标识适宜变速程度的标签的视频素材；
对所述样本视频片段提取视频帧，将所述视频帧确定为所述训练数据。
第五方面，根据本公开的一个或多个实施例，提供了一种电子设备，包括：至少一个处理器和存储器；
所述存储器存储计算机执行指令；
所述至少一个处理器执行所述存储器存储的计算机执行指令，使得所述至少一个处理器执行如上第一方面、第二方面以及第一方面、第二方面各种可能的设计所述的视频处理方法。
第六方面，根据本公开的一个或多个实施例，提供了一种计算机可读存储介质，所述计算机可读存储介质中存储有计算机执行指令，当处理器执行所述计算机执行指令时，实现如上第一方面、第二方面以及第一方面、第二方面各种可能的设计所述的视频处理方法。
第七方面，根据本公开的一个或多个实施例，提供一种计算机程序产品，包括计算机指令，该计算机指令被处理器执行时实现如上第一方面、第二方面以及第一方面、第二方面各种可能的设计所述的视频处理方法。
第八方面，根据本公开的一个或多个实施例，提供了一种计算机程序，当处理器执行计算机程序时，实现如上第一方面、第二方面以及第一方面、第二方面各种可能的设计所述的视频处理方法。
以上描述仅为本公开的较佳实施例以及对所运用技术原理的说明。本领域技术人员应当理解，本公开中所涉及的公开范围，并不限于上述技术特征的特定组合而成的技术方案，同时也应涵盖在不脱离上述公开构思的情况下，由上述技术特征或其等同特征进行任意组合而形成的其它技术方案。例如上述特征与本公开中公开的(但不限于)具有类似功能的技术特征进行互相替换而形成的技术方案。
此外，虽然采用特定次序描绘了各操作，但是这不应当理解为要求这些操作以所示出的特定次序或以顺序次序执行来执行。在一定环境下，多任务和并行处理可能是有利的。同样地，虽然在上面论述中包含了若干具体实现细节，但是这些不应当被解释为对本公开的范围的限制。在单独的实施例的上下文中描述的某些特征还可以组合地实现在单个实施例中。相反地，在单个实施例的上下文中描述的各种特征也可以单独地或以任何合适的子组合的方式实现在多个实施例中。
尽管已经采用特定于结构特征和/或方法逻辑动作的语言描述了本主题，但是应当理解所附权利要求书中所限定的主题未必局限于上面描述的特定特征或动作。相反，上面所描述的特定特征和动作仅仅是实现权利要求书的示例形式。

Claims (15)

  1. 一种视频处理方法，其特征在于，包括：
    获取初始视频片段；
    将所述初始视频片段输入机器学习模型中，并根据所述机器学习模型的输出结果确定所述初始视频片段的变速倍率；其中，所述机器学习模型已基于已标注了变速倍率的样本视频片段进行了训练；
    根据所述变速倍率对所述初始视频片段进行变速，得到目标视频片段。
  2. 根据权利要求1所述的方法，其特征在于，所述获取初始视频片段，包括：
    获取初始视频；
    通过对所述初始视频进行分镜，得到所述初始视频片段，其中，所述初始视频片段为所述初始视频中的分镜片段。
  3. 根据权利要求2所述的方法，其特征在于，在所述得到目标视频片段之后，还包括：
    通过在所述初始视频中将所述初始视频片段替换为所述目标视频片段，得到目标视频。
  4. 根据权利要求2所述的方法，其特征在于，所述通过对所述初始视频进行分镜，得到所述初始视频片段，包括：
    对所述初始视频进行分镜，得到所述初始视频中的分镜片段；
    将所述分镜片段分割为多个视频片段，将其中任一视频片段确定为所述初始视频片段。
  5. 根据权利要求1所述的方法，其特征在于，所述将所述初始视频片段输入机器学习模型中，并根据所述机器学习模型的输出结果确定所述初始视频片段的变速倍率，包括：
    提取所述初始视频片段的视频帧；
    将所述初始视频片段的视频帧输入所述机器学习模型中，获取所述初始视频片段对应的评价参数，所述评价参数为用于衡量所述初始视频片段适宜变速程度的大小的参数；
    根据所述初始视频片段对应的评价参数、以及预设评价参数与变速倍率对照关系，确定所述初始视频片段的变速倍率。
  6. 根据权利要求5所述的方法，其特征在于，在所述获取所述初始视频片段对应的评价参数后，还包括：
    对所述初始视频片段对应的评价参数以及其相邻的其他初始视频片段对应的评价参数进行平滑处理，将平滑处理后的各评价参数确定为对应的各初始视频片段最终的评价参数；或者
    在所述确定所述初始视频片段的变速倍率后，还包括：
    对所述初始视频片段的变速倍率以及其相邻的其他初始视频片段的变速倍率进行平滑处理，将平滑处理后的各变速倍率确定为对应的各初始视频片段最终的变速倍率。
  7. 根据权利要求5或6所述的方法，其特征在于，在所述确定所述初始视频片段的变速倍率后，还包括：
    获取所述初始视频片段中的初始语速；
    根据所述初始语速以及所述初始视频片段的变速倍率，确定所述初始视频片段按照所述变速倍率播放时的目标语速；
    若所述目标语速超过预设语速阈值，则根据所述预设语速阈值调整所述变速倍率。
  8. 一种模型训练方法，其特征在于，包括：
    获取多个训练数据；所述训练数据包括已标注了变速倍率的样本视频片段；
    根据所述训练数据对机器学习模型的初始模型进行训练，得到训练后的机器学习模型。
  9. 根据权利要求8所述的方法，其特征在于，所述获取多个训练数据，包括：
    获取所述已标注了变速倍率的样本视频片段，所述样本视频片段为已被慢放或快放、且被标注有慢放标签或快放标签的视频素材，和/或，未被慢放或快放、且其各视频片段分别被标注有标识适宜变速程度的标签的视频素材；
    对所述样本视频片段提取视频帧，将所述视频帧确定为所述训练数据。
  10. 一种视频处理设备，其特征在于，包括：
    获取单元，用于获取初始视频片段；
    处理单元，用于将所述初始视频片段输入机器学习模型中，并根据所述机器学习模型的输出结果确定所述初始视频片段的变速倍率；其中，所述机器学习模型已基于已标注了变速倍率的样本视频片段进行了训练；
    变速单元，用于根据所述变速倍率对所述初始视频片段进行变速，得到目标视频片段。
  11. 一种模型训练设备，其特征在于，包括：
    获取单元，用于获取多个训练数据；所述训练数据包括已标注了变速倍率的样本视频片段；
    训练单元，用于根据所述训练数据对机器学习模型的初始模型进行训练，得到训练后的机器学习模型。
  12. 一种电子设备，其特征在于，包括：至少一个处理器和存储器；
    所述存储器存储计算机执行指令；
    所述至少一个处理器执行所述存储器存储的计算机执行指令，使得所述至少一个处理器执行如权利要求1-9任一项所述的方法。
  13. 一种计算机可读存储介质，其特征在于，所述计算机可读存储介质中存储有计算机执行指令，当处理器执行所述计算机执行指令时，实现如权利要求1-9任一项所述的方法。
  14. 一种计算机程序产品，包括计算机指令，其特征在于，该计算机指令被处理器执行时实现如权利要求1-9任一项所述的方法。
  15. 一种计算机程序，其特征在于，所述计算机程序被处理器执行时实现如权利要求1-9任一项所述的方法。
PCT/CN2022/096646 2021-06-01 2022-06-01 视频处理方法、设备、存储介质及计算机程序产品 Ceased WO2022253276A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US18/566,180 US12451161B2 (en) 2021-06-01 2022-06-01 Method, device, storage medium and computer program product for video processing

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110610952.5A CN115442661B (zh) 2021-06-01 2021-06-01 视频处理方法、设备、存储介质及计算机程序产品
CN202110610952.5 2021-06-01

Publications (1)

Publication Number Publication Date
WO2022253276A1 true WO2022253276A1 (zh) 2022-12-08

Family

ID=84272213

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/096646 Ceased WO2022253276A1 (zh) 2021-06-01 2022-06-01 视频处理方法、设备、存储介质及计算机程序产品

Country Status (3)

Country Link
US (1) US12451161B2 (zh)
CN (1) CN115442661B (zh)
WO (1) WO2022253276A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
FR3163521A1 (fr) * 2024-06-12 2025-12-19 Orange Procédé et dispositif de modification d’au moins une séquence d’un contenu audiovisuel.

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7127127B2 (en) * 2003-03-04 2006-10-24 Microsoft Corporation System and method for adaptive video fast forward using scene generative models
CN104506947A (zh) * 2014-12-24 2015-04-08 福州大学 一种基于语义内容的视频快进/快退速度自适应调整方法
CN104735385A (zh) * 2015-03-31 2015-06-24 小米科技有限责任公司 播放控制方法及装置、电子设备
CN110730387A (zh) * 2019-11-13 2020-01-24 腾讯科技（深圳）有限公司 视频播放控制方法和装置、存储介质及电子装置
CN111771384A (zh) * 2018-02-28 2020-10-13 谷歌有限责任公司 自动调整回放速度和场境信息

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9502073B2 (en) * 2010-03-08 2016-11-22 Magisto Ltd. System and method for semi-automatic video editing
CN102271280A (zh) * 2011-07-20 2011-12-07 宝利微电子系统控股公司 一种数字音视频变速播放的方法和装置
CN108184169B (zh) * 2017-12-28 2020-10-09 Oppo广东移动通信有限公司 视频播放方法、装置、存储介质及电子设备
CN108401193A (zh) * 2018-03-21 2018-08-14 北京奇艺世纪科技有限公司 一种视频播放方法、装置和电子设备
CN109451248B (zh) 2018-11-23 2020-12-22 广州酷狗计算机科技有限公司 视频数据的处理方法、装置、终端及存储介质
CN110856042B (zh) * 2019-11-18 2022-02-11 腾讯科技（深圳）有限公司 视频播放方法、装置、计算机可读存储介质和计算机设备
CN112653920B (zh) * 2020-12-18 2022-05-24 北京字跳网络技术有限公司 视频处理方法、装置、设备及存储介质
CN112822546A (zh) * 2020-12-30 2021-05-18 珠海极海半导体有限公司 基于内容感知的倍速播放方法、系统、存储介质和设备

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7127127B2 (en) * 2003-03-04 2006-10-24 Microsoft Corporation System and method for adaptive video fast forward using scene generative models
CN104506947A (zh) * 2014-12-24 2015-04-08 福州大学 一种基于语义内容的视频快进/快退速度自适应调整方法
CN104735385A (zh) * 2015-03-31 2015-06-24 小米科技有限责任公司 播放控制方法及装置、电子设备
CN111771384A (zh) * 2018-02-28 2020-10-13 谷歌有限责任公司 自动调整回放速度和场境信息
CN110730387A (zh) * 2019-11-13 2020-01-24 腾讯科技（深圳）有限公司 视频播放控制方法和装置、存储介质及电子装置

Also Published As

Publication number Publication date
CN115442661A (zh) 2022-12-06
US20240249752A1 (en) 2024-07-25
CN115442661B (zh) 2024-03-19
US12451161B2 (en) 2025-10-21

Similar Documents

Publication Publication Date Title
JP7700285B2 (ja) ビデオ再生方法、装置、機器及び記憶媒体
WO2024012255A1 (zh) 语义分割模型训练方法、装置、电子设备及存储介质
CN109474850B (zh) 运动像素视频特效添加方法、装置、终端设备及存储介质
CN113259761B (zh) 视频处理方法和视频处理的设备、存储介质
JP7717852B2 (ja) 情報処理方法、装置、機器、可読記憶媒体及び製品
WO2023029904A1 (zh) 文本内容匹配方法、装置、电子设备及存储介质
WO2023284437A1 (zh) 媒体文件处理方法、装置、设备、可读存储介质及产品
CN117956100A (zh) 多媒体数据处理方法、装置、设备及介质
CN115981769A (zh) 页面显示方法、装置、设备、计算机可读存储介质及产品
CN112785670A (zh) 一种形象合成方法、装置、设备及存储介质
CN116994079B (zh) 图像处理方法及设备
CN114666666A (zh) 视频跳转播放方法、装置、终端设备及存储介质
WO2023000782A1 (zh) 获取视频热点的方法、装置、可读介质和电子设备
US20230419663A1 (en) Systems and Methods for Video Genre Classification
CN116931791A (zh) 图像加载方法、装置、电子设备及存储介质
US20240163500A1 (en) Information display method, apparatus, electronic device and storage medium
WO2022253276A1 (zh) 视频处理方法、设备、存储介质及计算机程序产品
CN115802104A (zh) 视频跳转播放方法、装置、电子设备及存储介质
WO2024046284A1 (zh) 绘制动画生成方法、装置、设备、可读存储介质及产品
CN116055798A (zh) 视频处理方法、装置及电子设备
CN114117127A (zh) 视频生成方法、装置、可读介质及电子设备
US12051445B1 (en) Method and apparatus of video editing, and electronic device and storage medium
CN117456063A (zh) 基于语音的人脸驱动方法、装置、电子设备及存储介质
CN113031775B (zh) 手势数据获取方法、装置、终端及存储介质
WO2023174066A1 (zh) 视频生成方法、装置、电子设备和存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22815318

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 18566180

Country of ref document: US

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 27.03.2024)

122 Ep: pct application non-entry in european phase

Ref document number: 22815318

Country of ref document: EP

Kind code of ref document: A1

WWG Wipo information: grant in national office

Ref document number: 18566180

Country of ref document: US