Method and device for acquiring high-energy fragment in game live broadcast

Abstract

The embodiment of the invention provides a method for obtaining a high-energy fragment in game live broadcast, and the method comprises the steps: carrying out the frame extraction processing of a game live broadcast video, so as to obtain a multi-frame game picture of the game live broadcast video; acquiring the game performance and the game time point of each frame of game picture in the multiple frames of game pictures; under the condition that the game performance changes, setting a game time point corresponding to the game performance as a high-energy moment; under the condition that the interval between the at least two adjacent high-energy moments is smaller than a first preset duration, determining the first high-energy moment and the last high-energy moment in the at least two adjacent high-energy moments; and editing the game live video between the first high-energy moment and the last high-energy moment into a high-energy fragment. According to the technical scheme provided by the embodiment of the invention, the starting position and the ending position of the high-energy fragment can be more accurately determined, so that the whole process of the high-energy fragment can be more accurately reflected.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of live broadcast technology, and in particular, to a method, apparatus, computer equipment, computer-readable storage medium, and computer program product for obtaining high-energy clips in a live game broadcast. Background Art

[0002] With the development of online interactive platforms, more and more users can communicate by logging into online interactive platforms. Among them, online live broadcasting is one of the popular online interactive platforms. Online live broadcasting can be game live broadcasting, entertainment live broadcasting or education live broadcasting, etc.

[0003] In game live broadcast, the anchor broadcasts the game process, such as League of Legends. The anchor usually hopes to automatically obtain the high-energy clips of the game live broadcast after the live broadcast, such as the clips of killing the opponent multiple times in a row, so as to deliver subsequent content.

[0004] However, the existing solutions for obtaining high-energy segments of games determine high-energy segments based on the broadcast of high-energy events in the game, and it is difficult to accurately obtain the start and end positions of the high-energy segments.

[0005] It should be noted that the above content is not necessarily prior art, nor is it intended to limit the scope of patent protection of this application. Summary of the invention

[0006] The embodiments of the present application provide a method, apparatus, computer device, computer-readable storage medium, and computer program product for obtaining high-energy clips in a live game broadcast, so as to solve or alleviate one or more of the technical problems raised above.

[0007] One aspect of an embodiment of the present application provides a method for obtaining high-energy clips in a live game broadcast, the method comprising: Performing frame extraction processing on the live game video to obtain multiple frames of game screens of the live game video; Obtaining the game record and game time point of each frame of the game screen in the multiple frames of the game screen; and When the game record changes, the game time point corresponding to the game record is set as a high-energy moment; In a case where an interval between at least two adjacent high-energy moments is less than a first preset duration, determining a first high-energy moment and a last high-energy moment of the at least two adjacent high-energy moments; The live game video between the first high-energy moment and the last high-energy moment is edited into high-energy clips.

[0008] Optionally, it also includes: When the multiple frames of game screen are acquired, each frame of the multiple frames of game screen is denoised.

[0009] Optionally, the method further comprises: Performing target detection on each frame of the game screen to identify and locate various information layouts of the game screen; Among them, the various information layouts of each frame of the game screen at least include a record column and a time column.

[0010] Optionally include: Text recognition is performed on various information layouts of each frame of the game screen to obtain the game record and the game time point corresponding to each frame of the game screen.

[0011] Optionally, the method further comprises: Merge the game record and game time information corresponding to each frame of the game screen into one text message; According to the time sequence of each frame of the multiple frames of game screen, the multiple text information are merged into an information table.

[0012] Optionally, the method further comprises: Traversing the information table; When it is detected that the game record in the text information is a preset initial value, the text information is set as a division point to distinguish different game sessions.

[0013] Optionally, the method further comprises: Setting a segment within a second preset time length before a first high-energy moment in the high-energy segment as a first supplementary segment; Setting a segment within a third preset time length after the last high-energy moment in the high-energy segment as a second supplementary segment; The first supplementary segment is added to the beginning of the high-energy segment, and the second supplementary segment is added to the end of the high-energy segment.

[0014] Another aspect of an embodiment of the present application provides a device for obtaining high-energy clips in a live game broadcast, the device comprising: A processing module, used for performing frame extraction processing on the live game video to obtain multiple frames of game screens of the live game video; An acquisition module, used to acquire the game record and game time point of each frame of the game screen in the multiple frames of the game screen; and A setting module, used for setting the game time point corresponding to the game record as a high-energy moment when the game record changes; A determination module, configured to determine a first high-energy moment and a last high-energy moment of the at least two adjacent high-energy moments when an interval between the at least two adjacent high-energy moments is less than a first preset duration; The editing module is used to edit the live game video between the first high-energy moment and the last high-energy moment into high-energy clips.

[0015] Another aspect of an embodiment of the present application provides a computer device, including: at least one processor; and a memory communicatively coupled to the at least one processor; Wherein: the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method as described above.

[0016] Another aspect of an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a processor, the method described above is implemented.

[0017] Another aspect of an embodiment of the present application provides a computer program product, including a computer program, which implements the method described above when executed by a processor.

[0018] The above technical solution adopted in the embodiment of the present application may have the following advantages: By detecting the changing moments of the game results of multiple frames of game screens in the live game video, combined with the corresponding game time points, the high-energy moments in the game process can be more accurately located. In addition, by setting a first preset duration, when the interval between at least two adjacent high-energy moments is less than the first preset duration, this solution determines the first high-energy moment and the last high-energy moment of at least two adjacent high-energy moments, and edits the live game video between the first high-energy moment and the last high-energy moment into a high-energy segment, so as to more accurately determine the starting position and the ending position of the high-energy segment, thereby more accurately reflecting the entire process of the high-energy segment. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The accompanying drawings exemplarily illustrate the embodiments and constitute a part of the specification, and together with the text description of the specification, are used to explain the exemplary implementation of the embodiments. The embodiments shown are for illustrative purposes only and do not limit the scope of the claims. In all drawings, the same reference numerals refer to similar but not necessarily identical elements.

[0020] Figure 1A diagram schematically shows an operating environment of a method for obtaining high-energy clips in a live game broadcast according to Embodiment 1 of the present application; Figure 2 A flowchart of a method for obtaining high-energy clips in a live game broadcast according to Embodiment 1 of the present application is schematically shown; Figure 3 A newly added flow chart of the method for obtaining high-energy clips in a live game broadcast according to the first embodiment of the present application is schematically shown; Figure 4 Another newly added flow chart of the method for obtaining high-energy clips in a live game broadcast according to the first embodiment of the present application is schematically shown; Figure 5 A further newly added flow chart of the method for obtaining high-energy clips in a live game broadcast according to the first embodiment of the present application is schematically shown; Figure 6 A diagram schematically shows an application example of the method for obtaining high-energy clips in a live game broadcast according to the first embodiment of the present application; Figure 7 A block diagram of a device for obtaining high-energy clips in a live game broadcast according to Embodiment 2 of the present application is schematically shown; and Figure 8 The hardware architecture diagram of the computer device according to the third embodiment of the present application is schematically shown. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solutions and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application.

[0022] It should be noted that the descriptions involving "first", "second", etc. in the embodiments of the present application are only for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0023] In the description of the present application, it should be understood that the numerical labels before the steps do not indicate the order in which the steps are executed, but are only used to facilitate the description of the present application and to distinguish each step, and therefore should not be understood as a limitation on the present application.

[0024] First, the following terms are explained: Frame extraction: A method of selecting specific frames from a video to reduce the amount of video data and retain key picture information. Through frame extraction, a certain number of key frames can be extracted from a game live broadcast or video for subsequent analysis and processing.

[0025] Game performance: refers to the performance and results of players in the game, usually measured by data such as kills, deaths, assists, etc. In games such as "League of Legends", performance is often expressed as "KDA" (kills, deaths, assists), reflecting the player's output and survivability.

[0026] Editing: refers to the process of selecting, cutting and splicing specific clips in a video, with the purpose of producing a complete video that meets the requirements. Through editing, unnecessary parts can be removed, key content can be highlighted or the order of the video can be adjusted to make it more compact and coherent, and easier to watch or spread.

[0027] Secondly, in order to facilitate those skilled in the art to understand the technical solutions provided in the embodiments of the present application, the relevant technologies are described below: With the development of online interactive platforms, more and more users can communicate by logging into online interactive platforms. Among them, online live broadcasting is one of the popular online interactive platforms. Online live broadcasting can be game live broadcasting, entertainment live broadcasting or education live broadcasting, etc.

[0028] In game live broadcast, the anchor broadcasts the game process, such as League of Legends. The anchor usually hopes to automatically obtain the high-energy clips of the game live broadcast after the live broadcast, such as the clips of killing the opponent multiple times in a row, so as to deliver subsequent content.

[0029] However, the existing solutions for obtaining high-energy segments of games determine high-energy segments based on the broadcast of high-energy events in the game, and it is difficult to accurately obtain the start and end positions of the high-energy segments.

[0030] To this end, an embodiment of the present application provides a technical solution for obtaining high-energy clips in a live game. In this technical solution, by detecting the moment when the game record of multiple frames of game screens in the live game video changes, combined with the corresponding game time points, the high-energy moments in the game process can be located more accurately. In addition, this solution sets a first preset duration, and when the interval between at least two adjacent high-energy moments is less than the first preset duration, determines the first high-energy moment and the last high-energy moment of at least two adjacent high-energy moments, and edits the live game video between the first high-energy moment and the last high-energy moment into a high-energy clip, so as to more accurately determine the starting position and the ending position of the high-energy clip, thereby more accurately reflecting the entire process of the high-energy clip. See below for details.

[0031] Finally, for ease of understanding, an exemplary operating environment is provided below.

[0032] like Figure 1 As shown, the operating environment diagram includes: Anchor terminal 2, server 4, and audience terminal (6A, 6B...6N). Anchor terminal 2 can upload live playback video data to server 4, and server 4 can provide high-energy clips edited from the live playback video data to anchor terminal 2 and audience terminal (6A, 6B...6N).

[0033] The server 4 can be a single server, a server cluster or a cloud computing service center. The server 4 can provide services such as storage, forwarding, configuration, and statistical analysis. For example, the live playback video data is analyzed and edited into multiple high-energy clips, and the multiple high-energy clips are forwarded to the anchor terminal 2 and the audience terminal (6A, 6B...6N).

[0034] The host terminal 2 can upload the live playback video file to the server 4 after the live broadcast ends to obtain the high-energy clips. The host terminal 2 can be an electronic device such as a smart phone, a tablet computer, or a virtual computing instance.

[0035] The audience terminals (6A, 6B...6N) may be any type of computer device, such as a smart phone, a tablet device, a laptop computer, a smart TV, a vehicle-mounted terminal, etc. The audience terminals (6A, 6B...6N) may have a built-in browser, a special program or a player for receiving data, such as obtaining a high-energy segment of a live playback video generated by the server 4.

[0036] The anchor terminal 2, the server 4, and the audience terminals (6A, 6B...6N) may be connected via a network. The network may include various network devices, such as routers, switches, multiplexers, hubs, modems, bridges, repeaters, firewalls, and / or proxy devices. The network may include physical links, such as coaxial cable links, twisted pair cable links, optical fiber links, combinations thereof, and / or the like. The network may include wireless links, such as cellular links, satellite links, Wi-Fi links, and / or the like.

[0037] It should be noted that the number of the host terminal 2, server 4, and viewer terminal in the figure is only for illustration and is not used to limit the scope of patent protection of this application. According to actual conditions, there can be any number of host terminals 2, servers 4, and viewer terminals.

[0038] The technical solutions of the present application are described below through multiple embodiments. It should be noted that these embodiments can be implemented in a variety of different forms and should not be construed as being limited to the embodiments described here.

[0039] Embodiment 1 Figure 2 The flowchart of the method for obtaining high-energy clips in a live game broadcast according to the first embodiment of the present application is schematically shown. Figure 2 As shown, the method for obtaining high-energy clips in a live game broadcast may include steps S200 to S208, wherein: Step S200, performing frame extraction processing on the live game video to obtain multiple frames of game screens of the live game video.

[0040] Step S202, obtaining the game record and game time point of each frame of the game screen in the multiple frames of the game screen.

[0041] Step S204, when the game record changes, the game time point corresponding to the game record is set as a high-energy moment.

[0042] Step S206, when the interval between at least two adjacent high-energy moments is less than the first preset duration, determine the first high-energy moment and the last high-energy moment of the at least two adjacent high-energy moments.

[0043] Step S208, editing the live game video between the first high-energy moment and the last high-energy moment into high-energy clips.

[0044] The method for obtaining high-energy clips in a live game provided in this embodiment can more accurately locate the high-energy moments in the game process by detecting the changing moments of the game scores of multiple frames of game screens in the live game video, combined with the corresponding game time points. In addition, this solution sets a first preset duration, and when the interval between at least two adjacent high-energy moments is less than the first preset duration, determines the first high-energy moment and the last high-energy moment of at least two adjacent high-energy moments, and edits the live game video between the first high-energy moment and the last high-energy moment into a high-energy clip, so as to more accurately determine the starting position and the ending position of the high-energy clip, thereby more accurately reflecting the entire process of the high-energy clip.

[0045] The following combination Figure 2 , each step in steps S200~S208 and other optional steps are explained in detail.

[0046] Step S200 , perform frame extraction processing on the game live video to obtain multiple frames of game screens of the game live video.

[0047] By performing frame extraction on live game videos, multiple frames of game images are extracted from the continuous video stream as basic data for subsequent analysis, which is used to further identify key moments or events in the game. By dividing the video stream into independent frames, changes in the game can be processed and analyzed frame by frame.

[0048] In one example, a certain time interval can be set, such as extracting one frame per second, to process the live game video in frames, which can ensure that each frame contains a complete game scene without increasing the processing burden of the system too much. Frames can be automatically extracted from the video stream at predetermined intervals to ensure that each frame represents a different moment in the live broadcast process.

[0049] For example, when the anchor is broadcasting the game of League of Legends, after the broadcast, the live video can be processed by setting a rule of extracting 1 frame per second to extract frames for the entire video. In a fierce team battle, the system will extract multiple key frames from this part of the video, and these frames will show different moments of the team battle, such as the release of players' skills, the killing of enemy characters, and changes in the battle situation.

[0050] Step S202 , obtain the game record and game time point of each frame of the game screen in the multiple frames of the game screen.

[0051] By obtaining the game record (such as kills, assists, deaths, etc.) and game time points of each frame in multiple game frames, each frame is associated with the corresponding game data to construct a time sequence of key events in the game. This step can provide the necessary information for the identification of subsequent high-energy moments, thereby accurately locating important game moments in complex game scenes.

[0052] In one example, the battle record data and game time data on the game screen can be identified through image recognition and text recognition technology. Specifically, each frame can be analyzed using image processing technology, and the battle record information and time information in the screen can be identified through algorithms. For example, edge detection, morphological processing and other methods can be applied to locate the area on the game interface that displays the battle record and time. Then, OCR (optical character recognition) technology can be combined to convert the text information in the screen into editable text data to identify battle record text and time text, such as "number of kills" and "number of assists".

[0053] In one example, a script automation tool can be used to simulate user operations, capture screens and locate coordinates, regularly capture images and analyze battle performance information. OCR technology can also be combined to extract information.

[0054] In one example, a machine learning model can be trained to identify the location and content of various battle record information in the game screen. Through deep learning technology, the model can learn the characteristics of the game interface and extract information more accurately.

[0055] Step S204 , when the game record changes, the game time point corresponding to the game record is set as a high-energy moment.

[0056] When the game record changes, the game time point corresponding to the change is set as a high-energy moment, which can accurately identify the time when the key event occurred during the game. By monitoring the changes in the record (such as kills, assists, etc.), it is possible to capture the moment when the player performs well in the game, providing accurate positioning for subsequent high-energy clip editing.

[0057] In one example, the game record may include the number of kills and the number of assists. A change in the game record may refer to a change in the number of kills, which means that the anchor has killed an enemy, and can be set as a high-energy moment. A change in the game record may also refer to a change in the number of assists, which means that the anchor is currently participating in a team battle and assisting teammates in killing, and can also be set as a high-energy moment.

[0058] In one example, a set of rules can be set to identify the conditions for changes in game performance, such as when a player's kill count increases, the number of deaths changes, or other key indicators change, the game performance can be considered to have changed. These indicators and the time points when they change can be monitored in real time and marked as high-energy moments.

[0059] For example, in a live broadcast of a League of Legends game, the host continuously kills enemy characters during the battle. During this process, it is monitored that the number of kills in the host's record suddenly increases from "2" to "3", and it occurs when the game time is "15:42", so the time point "15:42" is set as a high-energy moment. At the same time, if in the same game, the player successfully kills another enemy character at "16:10", and the number of kills changes again, the system will also set "16:10" as another high-energy moment. In this way, each of the host's high-energy moments can be accurately captured, providing a clear time node for the subsequent editing of high-energy clips.

[0060] Step S206 , when the interval between at least two adjacent high-energy moments is less than a first preset duration, determine the first high-energy moment and the last high-energy moment of the at least two adjacent high-energy moments.

[0061] When the interval between at least two adjacent high-energy moments is less than the first preset time length, the first and last high-energy moments are determined to accurately define the conditions for a kill streak. In games, the determination of a kill streak usually depends on a certain time interval. Only when the time interval is within the first preset time length can it be regarded as a continuous kill, and then multiple high-energy moments of continuous kills can be edited into a high-energy clip.

[0062] In one example, a suitable time threshold, such as 10 seconds, can be set as a criterion for determining whether adjacent high-energy moments constitute a kill streak. Each pair of adjacent high-energy moments can be compared, and if the time difference between them is less than the preset time length, it can be considered that these kills belong to the high-energy moments of the same kill streak event. The first high-energy moment can be marked as the starting point of the kill streak, and the last high-energy moment can be marked as the end point of the kill streak.

[0063] To illustrate by way of example, in a live broadcast of a game of "League of Legends", the anchor successfully killed the first enemy hero at "15:40", and the record was updated to "1". Then, at "15:41" and "15:42", the second and third enemy heroes were killed respectively, so that the record was updated to "2" and "3". Subsequently, at "15:50", the anchor killed another enemy, reaching "4". Since the time intervals between all high-energy moments are less than the first preset duration of 10 seconds, the anchor achieved a highlight performance of 4 kills. At this time, the first high-energy moment "15:40" can be determined as the first high-energy moment, and "15:50" can be determined as the last high-energy moment, and then the kill records between "15:40" and "15:50" can be merged into a continuous high-energy clip to show the anchor's wonderful performance.

[0064] Step S208 , the live game video between the first high-energy moment and the last high-energy moment is edited into a high-energy clip.

[0065] The game live video between the first high-energy moment and the last high-energy moment is edited into a high-energy clip to generate a video clip that can focus on the host's highlights. From the first high-energy moment to the last high-energy moment, the complete process from the beginning to the end of the high-energy event is covered, which can provide viewers with a smooth viewing experience and help the host to share or promote subsequent content.

[0066] In one example, an automated video editing algorithm can be used to take the time points of the first high-energy moment and the last high-energy moment as the start and end points of the editing. The frame number of each moment can be accurately located by reading the timestamp information of the live playback video file, and the screen from the first high-energy moment to the last high-energy moment can be cut. After the editing is completed, the clip can be marked as a "high-energy clip" for subsequent storage, display or sharing.

[0067] For example, it is detected that the anchor killed the first enemy at "15:40" (the first high-energy moment), and achieved a quad kill at "15:50" (the last high-energy moment), which means that the high-energy clip occurred between "15:40" and "15:50". Therefore, by editing the live playback video of this time period, all the killing scenes that occurred during this period can be merged into a high-energy clip. The edited high-energy clip shows the anchor's wonderful performance of killing 4 enemies in a row within 10 seconds. Saving this clip as a high-energy clip makes it convenient for the anchor or the audience to review and share this highlight moment after the live broadcast.

[0068] In this embodiment, by detecting the changing moments of the game scores of multiple frames of the game screen of the live game video, combined with the corresponding game time points, the high-energy moments in the game process can be more accurately located. In addition, by setting a first preset duration, when the interval between at least two adjacent high-energy moments is less than the first preset duration, this solution determines the first high-energy moment and the last high-energy moment of at least two adjacent high-energy moments, and edits the live game video between the first high-energy moment and the last high-energy moment into a high-energy segment, so as to more accurately determine the starting position and the ending position of the high-energy segment, thereby more accurately reflecting the entire process of the high-energy segment.

[0069] In an optional embodiment, the method for obtaining high-energy clips in a live game broadcast further includes: When the multiple frames of game screen are acquired, each frame of the multiple frames of game screen is denoised.

[0070] The acquired multiple frames of game images may have image quality problems due to network transmission, compression, etc. De-noising each frame can remove noise or interference information in the video image, thereby improving image clarity and quality, and ensuring that subsequent image analysis and information extraction are more accurate.

[0071] In one example, Gaussian filtering can be used for denoising. By weighted averaging the image, the noise area is smoothed while retaining the overall structure of the image. This is suitable for removing random noise in the video or slight blur caused by network transmission.

[0072] In one example, median filtering can be used for denoising. By replacing pixel values ​​with the median values ​​of the local area, the edge features of the image can be well preserved while removing obvious point noise.

[0073] In one example, bilateral filtering can be used for denoising. Bilateral filtering can not only remove noise but also retain edge information of the image while denoising. It is suitable for scenarios where it is necessary to reduce noise while retaining image details.

[0074] In this embodiment, by performing denoising on each frame of the multi-frame game screen, noise, blur and other interference information in the screen can be removed. This ensures the clarity and accuracy of the screen content, improves the accuracy of extracting the battle data and time point information in the live game video, and can more accurately identify and locate high-energy moments in subsequent processing, ensuring that the edited high-energy clips contain the correct battle changes and corresponding time points, and improves the positioning accuracy of the start and end positions of the high-energy clips.

[0075] In an optional embodiment, the method for obtaining high-energy clips in a live game broadcast further includes: Target detection is performed on each frame of the game screen to identify and locate various information layouts of the game screen, wherein the various information layouts of each frame of the game screen at least include a battle record column and a time column.

[0076] In one example, the various information layouts of the game screen may also include an in-game map thumbnail, current status information of a game character, a chat window, and a game character skill bar, etc.

[0077] In one example, an object detection algorithm, such as YOLO or Faster R-CNN, can be used to analyze each frame of a live game broadcast. The object detection algorithm can automatically draw bounding boxes for these areas in each frame and mark them so that data such as the score and time can be accurately extracted later.

[0078] In one example, a target detection algorithm model can be pre-trained using a large data set so that the model can be equipped with the ability to recognize different types of key information areas, such as the performance bar in the upper left corner of the screen and the time bar in the lower right corner.

[0079] As an example, the YOLO model is first pre-trained on a large data set to enable the model to recognize various layouts. Since the battle record bar and time bar in the game interface may not be within the recognition range of the pre-trained model, it is necessary to fine-tune the model or use a model designed specifically for the game interface. Each frame of the image is input into the trained YOLO model, and the model can output the predicted bounding box and category probability. According to the target detection results, the battle record bar in the upper left corner of the screen and the time bar in the lower right corner can be located, so that relevant information can be extracted from the identified battle record bar and time bar, such as the current number of kills, deaths, assists, and game time. Finally, this information can be saved together with the corresponding frame timestamp for subsequent analysis.

[0080] In this embodiment, target detection is performed on each frame of the game screen to identify and locate important information areas in the game screen, so that the key information areas can be accurately identified in the dynamically changing game scenes, reducing unnecessary information processing and laying the foundation for subsequent data extraction and analysis of high-energy moments.

[0081] In an optional embodiment, in step S202, "obtaining the game record and game time point of each frame of the game screen in the multiple frames of the game screen" includes: Text recognition is performed on various information layouts of each frame of the game screen to obtain the game record and the game time point corresponding to each frame of the game screen.

[0082] Since the acquired game screen is in image format, it is not convenient to process game-related information. Therefore, it is necessary to extract real-time updated battle record information and game time from the game screen to provide a data basis for subsequent analysis and editing of high-energy moments.

[0083] In one example, OCR (Optical Character Recognition) technology can be used to perform text recognition on the information layout of each frame of the game screen. The target detection algorithm identifies the bounding boxes of the record bar and the time bar, and then the OCR technology can scan the areas within these bounding boxes to extract the text information. For example, the record bar may display "5 / 1 / 2", indicating the number of kills, deaths, and assists of the player, while the time bar may display "12:34", indicating the time of the game. At this time, after OCR processing, the following information can be obtained: {"game_time": 12:34, "kda": [5,1,2]}.

[0084] In one example, the accuracy of text recognition can be enhanced by integrating a deep learning model. The system can use pre-trained OCR models (such as Tesseract or EasyOCR) for text recognition, and cooperate with image preprocessing techniques such as denoising and binarization to improve the recognition effect. For example, when extracting battle record information, the system first pre-processes the image of the battle record column to make the text clearer, and then uses OCR technology to extract text such as "kda:[5,1,2]".

[0085] In this embodiment, the layout of various types of information on the game screen is recognized through text, which provides necessary support for automatic content generation and data analysis. High-energy moments are automatically identified and extracted based on these data, thereby improving the efficiency and quality of content production.

[0086] In an optional embodiment, if Figure 3 As shown, the method for obtaining high-energy clips in a live game broadcast also includes: Step S300, combining the game results and game time information corresponding to each frame of the game screen into one text message.

[0087] Step S302: merging the multiple pieces of text information into an information table according to the time sequence of each frame of the multiple frames of game screen.

[0088] Since the text information of game records and game time points obtained by text recognition of various information layouts in each frame of the game screen is relatively scattered, multiple text information can be merged into an information table in chronological order, which can effectively manage and analyze the game live broadcast data and provide a clear timeline for subsequent high-energy moment extraction.

[0089] The time of each frame of the game screen refers to the time sequence of multiple frames of the game screen in the entire live playback video clip. The information tables are merged in chronological order, so that the high-energy moments can be edited and merged strictly according to the timeline of the live broadcast, which can more accurately display the start and end positions of the high-energy clips.

[0090] For example, if the current playback time of the video is 32 minutes and 09 seconds, then the timestamp of a frame of the game screen extracted at this second is 32:09. The timestamp indicates the relative position of the screen in the overall live broadcast, which is different from the concept of the game time point in the game match.

[0091] In one example, the information may be merged using a data structure such as a dictionary or JSON format.

[0092] In one example, the battle record and time point of each frame may be stored as a key-value pair. For example, the battle record "5 / 2 / 3" and the time "12:34" may be combined into "time: 12:34 - battle record: 5 / 2 / 3".

[0093] In one example, the merged information can be written to a database or file. For example, after each frame is processed, the merged information is recorded in the database to form a complete game data timeline. For example, the storage format can be "[time]: [game record]" to ensure fast retrieval and analysis when needed.

[0094] In one example, a programming language (such as Python) can be used to construct a dictionary or JSON format data structure to save relevant information at each time point. For example, the record and time point can be combined into a text message and stored in a dictionary, with a structure similar to: { "game_time":"32:09", "kda":[19,3,7] } Furthermore, the structure of merging multiple text messages into an information table may be similar to: [ { "game_time":"32:09", "kda":[19,3,7] " }, { "game_time":"33:15", "kda":[20,3,7] " } ] In this embodiment, by obtaining the information table merged in chronological order, the game data extracted from the live game video can be efficiently organized and utilized, which improves the readability of the data and provides a data basis for the subsequent analysis and processing of high-energy moments.

[0095] In an optional embodiment, if Figure 4 As shown, the method for obtaining high-energy clips in a live game broadcast also includes: Step S400, traversing the information table.

[0096] Step S402, when it is detected that the game record in the text information is a preset initial value, the text information is set as a division point to distinguish different game sessions.

[0097] Since the host may have played multiple games during the live broadcast of the game, it is necessary to identify and distinguish different game sessions when analyzing and processing the high-energy moments of the game, so that the key events and performances of each game can be processed independently, thereby improving the accuracy and effectiveness of data processing.

[0098] In one example, traversing the information table can be implemented by a loop structure of a programming language, such as a for loop or a while loop, and extracting relevant information by sequentially accessing each event in a list containing multiple events.

[0099] In one example, a specific game record initial value, such as "0 / 0 / 0", can be set to indicate the start of a new game. In the process of traversing the information table, if it is detected that the game record is the preset initial value, the text information at this time can be marked as a split point, thereby effectively distinguishing different game sessions.

[0100] In this embodiment, by using the text information of the preset initial value as the segmentation point to distinguish different game sessions, it is ensured that the information of different sessions can be clearly distinguished and processed, thereby effectively managing and organizing the game live broadcast data. Independent analysis of each different game can improve the accuracy and effectiveness of data processing.

[0101] In an optional embodiment, if Figure 5 As shown, the method for obtaining high-energy clips in a live game broadcast also includes: Step S500: setting a segment within a second preset time length before the first high-energy moment in the high-energy segment as a first supplementary segment.

[0102] Step S502: setting a segment within a third preset time length after the last high-energy moment in the high-energy segment as a second supplementary segment.

[0103] Step S504: adding the first supplementary segment to the beginning of the high-energy segment, and adding the second supplementary segment to the end of the high-energy segment.

[0104] In the editing of high-energy clips, only editing the high-energy clips themselves may not be able to fully show the whole process of the wonderful scene. In the short time before the high-energy clips, there may be some important background or pre-events happening. The supplement of the pre-event content plays an important role in the coherence and integrity of the whole clip.

[0105] In one example, the second preset duration can be set to 5 seconds. For example, if it is detected that the first high-energy moment in the high-energy clip to be edited occurs at "10:32", the system will automatically add the video clip from "10:27" to "10:32" to the front of the high-energy clip as the first supplementary clip. The high-energy clip with the first supplementary clip added not only includes the core content of the wonderful moment, but also adds enough background information to help restore more context, so that the audience can better understand the entire game scene when watching the high-energy clip, thereby improving the audience's viewing experience.

[0106] After a high-energy moment occurs, the exciting scene may not be completely over, especially in complex game scenes. There may still be subsequent developments or events after the last high-energy moment, which play an important role in summarizing or continuing the high-energy moment. By setting the third preset duration and adding the clips after the last high-energy moment, these subsequent key scenes can be retained, ensuring the integrity and coherence of the high-energy clips, thereby enhancing the audience's viewing experience.

[0107] In one example, the third preset duration can be set to 5 seconds. When the last high-energy moment is detected, such as a "triple kill" occurring at "12:45", the video clip from "12:45" to "12:50" can be automatically supplemented as the second supplementary clip. By adding the second supplementary clip, it is ensured that the content after the high-energy moment can also be edited in, capturing subsequent events that may occur, such as enemy escape, team advancement, etc.

[0108] In one example, the length or content of the supplementary clip can be adjusted according to the specific scene. For example, there may be a long preparation phase before some high-energy moments, such as ambushes and skill releases. In this case, the first supplementary clip can be set to be slightly longer, such as 8 to 10 seconds. After some high-energy moments, the battle situation may change quickly, and the second supplementary clip can also be shortened or extended accordingly. Dynamically adjusting the supplementary content according to different game scenes can ensure that the audience obtains more complete contextual information.

[0109] Exemplarily, when the anchor completes a triple kill between "13:15" and "13:20", "13:15" to "13:20" are edited into a high-energy clip. The last time point is marked as the last high-energy moment, and the first time point is marked as the first high-energy moment. In order to fully display the battle situation before and after the high-energy moment, two clips can be added at the first high-energy moment and the last high-energy moment, respectively. Specifically, the system adds the clip from "13:10" to "13:15" to the beginning of the high-energy clip to show the player's preparation stage before the high-energy moment, such as strategic operations such as positioning, skill charging, or ambushing before team battles. Through this supplement, the audience can better understand the cause of this battle and understand how players create opportunities to enter high-energy moments. Then, the clip between "13:20" and "13:25" is added to the end of the high-energy clip to show the evolution of the situation after the battle, such as the reaction of the enemy team, the advancement or retreat of one's own team, or the competition for map resources after the team battle.

[0110] In this embodiment, by performing supplementary processing at the beginning and end of the high-energy segment, the entire high-energy segment not only shows the core battle process, but also includes the foreshadowing before the battle and the ending after the battle. This enhances the continuity and fun of watching, ensures that the audience has a complete understanding of the overall picture of this battle, and enhances the audience's viewing experience.

[0111] In order to make this application easier to understand, the following Figure 6 An exemplary application is provided.

[0112] S11, obtain the live playback video, and perform frame extraction on the video. The FFmpeg tool can be used to extract one frame per second.

[0113] S12, performing denoising processing on each extracted frame.

[0114] S13, through target detection technology, identifies and locates various key layouts in the game screen, such as the time bar and the kill bar.

[0115] S14, using OCR technology to identify the located time bar and text box, and extracting the host's kill information and key data of the current game time.

[0116] S15, summarizing the acquired kill information and game time into an information list in chronological order.

[0117] S16, traverse the information list, detect that the record is reset to 0, and determine that a new game is started.

[0118] S17, traverse the information list, detect that the battle record has changed, and take the changed frame as a kill event.

[0119] S18, detecting the adjacent time intervals of each killing event, if the time intervals of multiple killing events are less than 10 seconds, then merging the multiple killing events into a high-energy segment, wherein the first killing event is the beginning of the high-energy segment, and the last killing event is the end of the high-energy segment.

[0120] S19, insert the first 5 seconds of video content of the first kill event at the beginning of the high-energy segment.

[0121] S20, inserting the last 5 seconds of video content of the last killing event at the end of the high-energy segment, thereby generating a final high-energy segment.

[0122] In this exemplary application, by detecting the changing moments of the game scores of multiple frames of game screens in the live game video, combined with the corresponding game time points, the high-energy moments in the game process can be more accurately located. In addition, this solution sets a first preset duration, and when the interval between at least two adjacent high-energy moments is less than the first preset duration, determines the first high-energy moment and the last high-energy moment of at least two adjacent high-energy moments, and edits the live game video between the first high-energy moment and the last high-energy moment into a high-energy segment, so as to more accurately determine the starting position and the ending position of the high-energy segment, thereby more accurately reflecting the entire process of the high-energy segment.

[0123] Embodiment 2 Figure 7 The block diagram of the device for obtaining high-energy clips in a live game broadcast according to the second embodiment of the present application is schematically shown. The device can be divided into one or more program modules, one or more program modules are stored in a storage medium, and are executed by one or more processors to complete the embodiment of the present application. The program module referred to in the embodiment of the present application refers to a series of computer program instruction segments that can complete specific functions. The following description will specifically introduce the functions of each program module in this embodiment. Figure 7 As shown, the device 1000 may include: a processing module 1100, an acquisition module 1200, a setting module 1300, a determination module 1400, and a clipping module 1500, wherein: The processing module 1100 is used to perform frame extraction processing on the live game video to obtain multiple frames of game screens of the live game video.

[0124] The acquisition module 1200 is used to acquire the game record and game time point of each frame of the game screen in the multiple frames of the game screen.

[0125] The setting module 1300 is used to set the game time point corresponding to the game record as a high-energy moment when the game record changes.

[0126] The determination module 1400 is used to determine the first high-energy moment and the last high-energy moment of the at least two adjacent high-energy moments when the interval between the at least two adjacent high-energy moments is less than a first preset duration.

[0127] The editing module 1500 is used to edit the live game video between the first high-energy moment and the last high-energy moment into high-energy clips.

[0128] As an optional embodiment, the device 1000 is further used for: When the multiple frames of game screen are acquired, each frame of the multiple frames of game screen is denoised.

[0129] As an optional embodiment, the device 1000 is further used for: Performing target detection on each frame of the game screen to identify and locate various information layouts of the game screen; Among them, the various information layouts of each frame of the game screen at least include a record column and a time column.

[0130] As an optional embodiment, the device 1000 is further used for: Text recognition is performed on various information layouts of each frame of the game screen to obtain the game record and the game time point corresponding to each frame of the game screen.

[0131] As an optional embodiment, the device 1000 is further used for: Merge the game record and game time information corresponding to each frame of the game screen into one text message; According to the time sequence of each frame of the multiple frames of game screen, the multiple text information are merged into an information table.

[0132] As an optional embodiment, the device 1000 is further used for: Traversing the information table; When it is detected that the game record in the text information is a preset initial value, the text information is set as a division point to distinguish different game sessions.

[0133] As an optional embodiment, the device 1000 is further used for: Setting a segment within a second preset time length before a first high-energy moment in the high-energy segment as a first supplementary segment; Setting a segment within a third preset time length after the last high-energy moment in the high-energy segment as a second supplementary segment; The first supplementary segment is added to the beginning of the high-energy segment, and the second supplementary segment is added to the end of the high-energy segment.

[0134] Embodiment 3 Figure 8 The schematic diagram of the hardware architecture of a computer device 10000 suitable for implementing the method for obtaining high-energy clips in a live game broadcast according to the third embodiment of the present application is shown schematically. In some embodiments, the computer device 10000 may be a terminal device such as a smart phone, a wearable device, a tablet computer, a personal computer, a vehicle-mounted terminal, a game console, a virtual device, a workbench, a digital assistant, a set-top box, a robot, etc. In other embodiments, the computer device 10000 may be a rack server, a blade server, a tower server, or a cabinet server (including an independent server, or a server cluster composed of multiple servers), etc. Figure 8 As shown, the computer device 10000 includes but is not limited to: a memory 10010, a processor 10020, and a network interface 10030 that can communicate with each other through a system bus. Among them: The memory 10010 includes at least one type of computer-readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (such as SD or DX memory), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 10010 can be an internal storage module of the computer device 10000, such as a hard disk or memory of the computer device 10000. In other embodiments, the memory 10010 can also be an external storage device of the computer device 10000, such as a plug-in hard disk equipped on the computer device 10000, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. Of course, the memory 10010 can also include both the internal storage module of the computer device 10000 and its external storage device. In this embodiment, the memory 10010 is generally used to store the operating system and various application software installed on the computer device 10000, such as the program code of the method for obtaining high-energy clips in the live game broadcast, etc. In addition, the memory 10010 can also be used to temporarily store various data that have been output or are to be output.

[0135] In some embodiments, the processor 10020 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other chips. The processor 10020 is generally used to control the overall operation of the computer device 10000, such as performing control and processing related to data interaction or communication with the computer device 10000. In this embodiment, the processor 10020 is used to run the program code stored in the memory 10010 or process data.

[0136] The network interface 10030 may include a wireless network interface or a wired network interface, and the network interface 10030 is generally used to establish a communication link between the computer device 10000 and other computer devices. For example, the network interface 10030 is used to connect the computer device 10000 to an external terminal through a network, and to establish a data transmission channel and a communication link between the computer device 10000 and the external terminal. The network may be a wireless or wired network such as an intranet, the Internet, the Global System of Mobile communication (GSM), Wideband Code Division Multiple Access (WCDMA), 4G network, 5G network, Bluetooth, Wi-Fi, etc.

[0137] It should be pointed out that Figure 8 Only a computer device having components 10010 - 10030 is shown, but it should be understood that implementation of all of the components shown is not a requirement, and more or fewer components may alternatively be implemented.

[0138] In this embodiment, the method for obtaining high-energy clips in game live broadcast stored in memory 10010 can also be divided into one or more program modules and executed by one or more processors (such as processor 10020) to complete the embodiment of the present application.

[0139] Embodiment 4 An embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored, wherein when the computer program is executed by a processor, the steps of the method for obtaining high-energy clips in a live game broadcast in the embodiment are implemented.

[0140] In this embodiment, the computer-readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the computer-readable storage medium may be an internal storage unit of a computer device, such as a hard disk or memory of the computer device. In other embodiments, the computer-readable storage medium may also be an external storage device of a computer device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the computer device. Of course, the computer-readable storage medium may also include both an internal storage unit of a computer device and an external storage device thereof. In this embodiment, the computer-readable storage medium is generally used to store an operating system and various application software installed on a computer device, such as the program code of a method for obtaining high-energy fragments in a live game broadcast in the embodiment. In addition, the computer-readable storage medium may also be used to temporarily store various types of data that have been output or are to be output.

[0141] Embodiment 5 An embodiment of the present application also provides a computer program product, including a computer program, which implements the method in the above embodiment when executed by a processor.

[0142] Obviously, those skilled in the art should understand that the modules or steps of the above-mentioned embodiments of the present application can be implemented by general-purpose computer devices, they can be concentrated on a single computer device, or distributed on a network composed of multiple computer devices, optionally, they can be implemented by executable program codes of computer devices, so that they can be stored in a storage device and executed by the computer device, and in some cases, the steps shown or described can be executed in a different order from that herein, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. In this way, the embodiments of the present application are not limited to any specific combination of hardware and software.

[0143] It should be noted that the above are only preferred embodiments of the present application, and the patent protection scope of the present application is not limited thereto. Any equivalent structure or equivalent process transformation made using the contents of the specification and drawings of the present application, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A method for obtaining high-energy clips in a live game broadcast, characterized in that: The method comprises: Performing frame extraction processing on the live game video to obtain multiple frames of game screens of the live game video; Obtaining the game record and game time point of each frame of the game screen in the multiple frames of the game screen; and When the game record changes, the game time point corresponding to the game record is set as a high-energy moment; In a case where an interval between at least two adjacent high-energy moments is less than a first preset duration, determining a first high-energy moment and a last high-energy moment of the at least two adjacent high-energy moments; The live game video between the first high-energy moment and the last high-energy moment is edited into high-energy clips.

2. The method according to claim 1, characterized in that The method further comprises: When the multiple frames of game screen are acquired, each frame of the multiple frames of game screen is denoised.

3. The method according to claim 2, characterized in that The method further comprises: Performing target detection on each frame of the game screen to identify and locate various information layouts of the game screen; Among them, the various information layouts of each frame of the game screen at least include a record column and a time column.

4. The method according to claim 3, characterized in that: Obtaining the game record and game time point of each frame of the game screen in the multiple frames of the game screen, including: Text recognition is performed on various information layouts of each frame of the game screen to obtain the game record and the game time point corresponding to each frame of the game screen.

5. The method according to claim 4, characterized in that The method further comprises: Merge the game record and game time information corresponding to each frame of the game screen into one text message; According to the time sequence of each frame of the multiple frames of game screen, the multiple text information are merged into an information table.

6. The method according to claim 5, characterized in that The method further comprises: Traversing the information table; When it is detected that the game record in the text information is a preset initial value, the text information is set as a dividing point to distinguish different game sessions.

7. The method according to any one of claims 1 to 6, characterized in that: The method further comprises: Setting a segment within a second preset time length before a first high-energy moment in the high-energy segment as a first supplementary segment; Setting a segment within a third preset time length after the last high-energy moment in the high-energy segment as a second supplementary segment; The first supplementary segment is added to the beginning of the high-energy segment, and the second supplementary segment is added to the end of the high-energy segment.

8. A device for obtaining high-energy clips in a live game broadcast, characterized in that: The device comprises: A processing module, used for performing frame extraction processing on the live game video to obtain multiple frames of game screens of the live game video; An acquisition module, used to acquire the game record and game time point of each frame of the game screen in the multiple frames of the game screen; and A setting module, used for setting the game time point corresponding to the game record as a high-energy moment when the game record changes; A determination module, configured to determine a first high-energy moment and a last high-energy moment of the at least two adjacent high-energy moments when an interval between the at least two adjacent high-energy moments is less than a first preset duration; The editing module is used to edit the live game video between the first high-energy moment and the last high-energy moment into high-energy clips.

9. A computer device, characterized in that: include: at least one processor; and a memory communicatively connected to the at least one processor; wherein: The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a processor, the method according to any one of claims 1 to 7 is implemented.

11. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to claims 1 to 7 are implemented.

Images