Video processing method, video playing method, device, equipment and storage medium

Abstract

The present disclosure relates to a video processing method, a video playing method, an apparatus, a device and a storage medium, obtaining a derivative plot text corresponding to at least one first playing position in a first video; obtaining event information and role information associated with the first playing position from a preset knowledge graph, the preset knowledge graph being determined in advance according to a plot logic of the first video; generating derivative guide information corresponding to the first playing position according to the event information, the role information and the derivative plot text; and generating a derivative video segment corresponding to the first playing position based on the plot logic, the derivative guide information, a reference segment of the first playing position and the derivative plot text, the derivative video segment not belonging to original video content of the first video. Thus, the event information and the role information are first determined based on the preset knowledge graph representing the plot logic, and then the derivative content conforming to the plot logic can be determined based on the above information, thereby improving the production quality and production efficiency of the derivative content.

Description

Technical Field

[0001] This disclosure relates to the field of information processing technology, and in particular to a video processing method, a video playback method, an apparatus, a device, and a storage medium. Background Technology

[0002] With the rapid development of video technology, watching videos has become a popular way for users to relax and unwind. While watching videos, users often want to see related content (such as spin-offs). Therefore, generating derivative content for videos and presenting it to users during playback can enhance their video viewing experience.

[0003] In related technologies, derivative content of videos is generally determined manually, or generated using large models. However, relying on manual methods requires high labor costs, and existing large models cannot follow the plot logic of videos, resulting in the predicted derivative content often contradicting the logic of the video. Summary of the Invention

[0004] To address the aforementioned technical problems, this disclosure provides a video processing method, a video playback method, an apparatus, a device, and a storage medium.

[0005] Firstly, this disclosure provides a video processing method, including: Obtain the derived plot text corresponding to at least one first playback position in the first video; The event information and character information associated with the first playback position are obtained from the preset knowledge graph corresponding to the first video, wherein the preset knowledge graph is predetermined according to the plot logic of the first video; Based on the event information, the character information, and the derivative plot text, generate derivative guidance information corresponding to the first playback position; Based on the plot logic, the derived guidance information, the reference segment at the first playback position, and the derived plot text, a derived video segment corresponding to the first playback position is generated. The derived video segment does not belong to the original video content of the first video. The reference segment includes the first video content of a first time period before the first playback position and / or the second video content of a second time period after the first playback position.

[0006] Secondly, this disclosure provides a video playback method, including: When the first video is played to the first playback position, information related to at least one derivative video segment associated with the first playback position is displayed, the derivative video segment being obtained based on the method provided in the first aspect above; In response to a selection operation of a first relevant information among at least one of the relevant information, a first derived video segment corresponding to the first relevant information is played.

[0007] Thirdly, this disclosure provides a video processing apparatus, including: The first acquisition module is used to acquire derivative plot text corresponding to at least one first playback position in the first video; The second acquisition module is used to acquire event information and character information associated with the first playback position from the preset knowledge graph corresponding to the first video, wherein the preset knowledge graph is predetermined according to the plot logic of the first video; The first generation module is used to generate derivative guidance information corresponding to the first playback position based on the event information, the character information, and the derivative plot text. The second generation module is used to generate a derivative video segment corresponding to the first playback position based on the plot logic, the derivative guidance information, the reference segment of the first playback position, and the derivative plot text. The derivative video segment does not belong to the original video content of the first video. The reference segment includes the first video content of the first time period before the first playback position and / or the second video content of the second time period after the first playback position.

[0008] Fourthly, this disclosure provides a video playback device, including: The display module is configured to display relevant information of at least one derivative video segment associated with the first playback position when the first video is played to the first playback position, wherein the derivative video segment is obtained based on the method provided in the first aspect above; The first playback module is configured to play a first derived video segment corresponding to the first relevant information in response to a selection operation of at least one of the relevant information.

[0009] Fifthly, embodiments of this disclosure also provide an electronic device, including: One or more processors; Storage device for storing one or more programs. When one or more programs are executed by one or more processors, the one or more processors implement the methods provided by the first or second aspect.

[0010] In a sixth aspect, embodiments of this disclosure also provide a computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the method provided in the first or second aspect.

[0011] The technical solution provided in this disclosure has the following advantages compared with the prior art: This disclosure discloses a video processing method, video playback method, apparatus, device, and storage medium that obtains derived plot text corresponding to at least one first playback position in a first video; obtains event information and character information associated with the first playback position from a preset knowledge graph corresponding to the first video, the preset knowledge graph being predetermined according to the plot logic of the first video; generates derived guidance information corresponding to the first playback position based on the event information, character information, and derived plot text; and generates a derived video segment corresponding to the first playback position based on the plot logic, derived guidance information, reference segments of the first playback position, and derived plot text. The derived video segment is not part of the original video content of the first video, and the reference segments include the first video content of a first time period before the first playback position and / or the second video content of a second time period after the first playback position. Thus, by first determining event information and character information based on a preset knowledge graph representing the plot logic, then generating guidance text for the derived video segment based on the event information, character information, and derived plot text, and combining the guidance text, plot logic, reference materials, and derived plot text, derived content conforming to the plot logic is generated, improving the production quality and efficiency of derived content. Attached Figure Description

[0012] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.

[0013] To more clearly illustrate the technical solutions in the embodiments of this disclosure or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 A flowchart illustrating a video processing method provided in an embodiment of this disclosure; Figure 2 A flowchart illustrating another video processing method provided in this embodiment of the present disclosure; Figure 3 A flowchart illustrating yet another video processing method provided in this disclosure embodiment; Figure 4 A flowchart illustrating a video playback method provided in an embodiment of this disclosure; Figure 5 A logical schematic diagram of a video processing method and a video playback method provided in the embodiments of this disclosure; Figure 6 This is a schematic diagram of the structure of a video processing apparatus provided in an embodiment of the present disclosure; Figure 7This is a schematic diagram of the structure of a video playback device provided in an embodiment of the present disclosure; Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure. Detailed Implementation

[0015] for To better understand the above-mentioned objectives, features, and advantages of this disclosure, the solutions disclosed herein will be further described below. It should be noted that, unless otherwise specified, the embodiments and features described herein can be combined with each other.

[0016] Numerous specific details are set forth in the following description in order to provide a full understanding of this disclosure, but this disclosure may also be implemented in other ways different from those described herein; obviously, the embodiments in the specification are only some, and not all, of the embodiments of this disclosure.

[0017] During the viewing process of movies and TV dramas, there are problems such as the limited forms of interaction between the movies and TV dramas and the users, the disconnect between the video processing process and the video creation process, and the separation between creation and user needs.

[0018] Among these issues, the interaction between film and television dramas and users is limited to a single form, specifically manifested in the fact that users are mainly passively watching, and user interaction is mostly limited to shallow feedback such as comments and bullet screens. This fails to establish a closed loop of "user prediction - content generation - creative feedback," and user input is only used as data collection and cannot be transformed into personalized content in real time, resulting in a lack of user participation in the plot.

[0019] In particular, the video processing process is disconnected from the video creation process. Specifically, the video is pushed to users based on user voting and fixed branching, which fails to generate derivative content that is consistent with the plot logic, resulting in the derivative content of the video being disconnected from the original plot.

[0020] Among these issues, the disconnect between content creation and user needs manifests itself in the following ways: film and television producers struggle to capture user preferences, and point rewards are often linked to consumption behaviors (such as recharging and watching time), failing to deeply integrate with content interaction. This leads to video creation relying on experience-based judgment, resulting in biases and risks.

[0021] To solve the above problems, the following will combine... Figures 1-3 The video processing method provided in the embodiments of this disclosure will be described. In the embodiments of this disclosure, the video processing method can be executed by a video processing device. The video processing device can be an electronic device or a server.

[0022] Figure 1 A flowchart illustrating a video processing method provided in an embodiment of this disclosure is shown.

[0023] like Figure 1 As shown, the video processing method may include the following steps.

[0024] S110. Obtain the derived plot text corresponding to at least one first playback position in the first video.

[0025] The first video is a film or television drama that needs to be processed to generate derivative content.

[0026] The first playback position is the start time of the playback of the derivative content. Optionally, the first video may include one or more first playback positions, such that the first video is associated with derivative content at one or more first playback positions.

[0027] Among them, the derivative plot text is the script information for generating derivative content for the first video.

[0028] It should be noted that derivative story texts are not part of the original story text of the first video. For example, derivative story texts are script information for spin-off dramas of the first video.

[0029] S120. Obtain event information and character information associated with the first playback position from the preset knowledge graph corresponding to the first video, wherein the preset knowledge graph is predetermined according to the plot logic of the first video.

[0030] The plot logic of the first video refers to the global constraint information of the first video. In other words, the plot logic of the first video comprises constraints covering the entire plot of the first video. Optionally, the plot logic includes at least one of the following: event causal constraints, character personality constraints, and style characteristic constraints.

[0031] Understandably, each playback position of the first video is associated with corresponding event information and role information. For the first playback position, the video processing device uses the first playback position as an index to search for the corresponding event information and role information from the preset knowledge graph.

[0032] Optionally, event information includes, but is not limited to, event name and event description. The event description includes the event's attribute information, including but not limited to the location, scenario, and time of occurrence.

[0033] Optionally, character information includes, but is not limited to, character name and character description. The character description is the character's attribute information, including but not limited to the character's personality traits, role, and nickname.

[0034] In this way, for any first playback position, key entities can be extracted from the preset knowledge graph determined by the plot logic of the first video, thereby accurately determining the event information and character information corresponding to that first playback position.

[0035] S130. Based on event information, character information, and derivative plot text, generate derivative guidance information corresponding to the first playback position.

[0036] Among them, the derivative guidance information is the guiding text used to generate derivative content, which can serve as real-time constraint information for generating derivative content. Optionally, the derivative guidance information can be represented in the form of a demand vector.

[0037] In some embodiments, S130 is specifically implemented through the following steps: determining the derivative plot type corresponding to the derivative plot text; and splicing the event information, character information, and derivative plot type according to a preset splicing format to generate derivative guidance information.

[0038] Among them, the derivative plot type refers to the category of derivative content. Optionally, the derivative plot type can be either behavior-based or outcome-based.

[0039] For example, if the derivative plot type is "the protagonist chooses option A", then the derivative plot type is an action type.

[0040] For example, if the derivative plot type is "Will the protagonist end up with the second male lead?", then the derivative plot type is a result type.

[0041] The preset splicing format is the expression form of the derived guiding information. Optionally, when the preset splicing format is represented in vector form, the preset splicing format can be represented as: <character, derived plot type, {character description, event and event description}>.

[0042] Specifically, the video processing device first inputs the derived plot text into a preset plot classification model for classification to obtain the derived plot type corresponding to the derived plot text; then, the video processing device splices the events and their corresponding event descriptions in the event information, the characters and their corresponding character descriptions in the character information, and the derived plot type according to a preset splicing format to generate derived guidance information.

[0043] The preset plot classification model can be a fine-tuned model used for information classification. Optionally, the preset plot classification model includes, but is not limited to, the BERT model.

[0044] For example, the first video is "XX", and the first playback position is the 20-minute mark of episode 10 of "XX". The video processing device obtains the derivative plot text corresponding to the 20-minute mark of episode 10 of "XX". This derivative plot text describes the derivative plot of "the protagonist chooses option A". Then, the video processing device can generate the corresponding introductory text for the 20-minute mark of episode 10 of "XX" based on the derivative plot of "the protagonist chooses option A", event information and character information.

[0045] In this way, converting the information obtained above into vector form can simplify the information expression, and determine the introductory text used to generate derivative content based on this simplified information.

[0046] S140. Based on the plot logic, derivative guiding information, reference segments at the first playback position, and derivative plot text, generate derivative video segments corresponding to the first playback position.

[0047] Derivative video clips are content derived from the first playback position. Specifically, derivative video clips are not part of the original video content of the first video. For example, a derivative video clip might be a spin-off of the first video.

[0048] The reference segment is a reference material used to generate a derivative video segment at the first playback position. The video content of the reference segment can be used to constrain the event logic of the derivative content.

[0049] Optionally, the reference segment includes the first video content of a first time period before the first playback position and / or the second video content of a second time period after the first playback position.

[0050] In one scenario, the first and second time periods are the same, for example, both are 30 minutes long. In another scenario, the first and second time periods are different, for example, the first time period is 30 minutes and the second time period is 10 minutes.

[0051] In some embodiments, the video processing device can use the plot logic as global constraint information and the derived guidance information as real-time constraint information. It inputs the reference segment and derived plot text of the first playback position into a preset large model, and performs video segment prediction according to the global constraint information and real-time constraint information to obtain the derived video segment corresponding to the first playback position.

[0052] In this way, by combining introductory texts, plot logic, reference materials, and derivative plot texts, it is possible to generate derivative content that conforms to the plot logic, thereby improving the quality of derivative content.

[0053] This disclosure discloses a video processing method that involves: obtaining derived plot text corresponding to at least one first playback position in a first video; obtaining event information and character information associated with the first playback position from a preset knowledge graph corresponding to the first video, wherein the preset knowledge graph is predetermined according to the plot logic of the first video; generating derived guidance information corresponding to the first playback position based on the event information, character information, and derived plot text; and generating a derived video segment corresponding to the first playback position based on the plot logic, derived guidance information, reference segments of the first playback position, and derived plot text. The derived video segment is not part of the original video content of the first video, and the reference segments include the first video content of a first time period before the first playback position and / or the second video content of a second time period after the first playback position. Thus, by first determining event information and character information based on a preset knowledge graph representing the plot logic, then generating guidance text for the derived video segment based on the event information, character information, and derived plot text, and combining the guidance text, plot logic, reference materials, and derived plot text, derived content conforming to the plot logic is generated, thereby improving the production quality and efficiency of derived content.

[0054] In another embodiment of this application, after performing S140, the video processing method further includes heat calculation, closed-loop feedback, and points distribution process.

[0055] In this embodiment, after executing S140, the video processing method further includes the following steps: obtaining voting data, viewing completion rate, and positive comment data for the derivative video clip; calculating the popularity value of the derivative video clip based on the voting data, viewing completion rate, and positive comment data; if the popularity value of the derivative video clip is greater than the popularity value of other derivative content, then continuing to process the first video based on the derivative video clip and the positive comment data of the derivative video clip; and determining the points to be awarded to the viewing user based on the popularity value of the derivative video clip.

[0056] Voting data can be represented by the number of votes cast by users for derivative video clips. Completion rate can be represented by the viewing time of users for derivative video clips. Positive comment data can be represented by positive comment keywords from users for derivative video clips.

[0057] Optionally, the popularity value of the derived video clip can be determined by the following method: weighting and summing the voting data, the first weight corresponding to the voting data, the viewing completion rate, the second weight corresponding to the viewing completion rate, the positive comment data, and the third weight corresponding to the positive comment data to obtain the popularity value of the derived video clip, wherein the sum of the first weight, the second weight, and the third weight is 1.

[0058] For example, the first weight is 0.6, the second weight is 0.2, and the third weight is 0.2.

[0059] The points awarded to viewers are determined based on a base score and a contribution score corresponding to the popularity value.

[0060] Optionally, the base score is a preset fixed value, such as 10 points. The contribution score corresponding to the popularity score can be obtained by multiplying the popularity score by a preset ratio. The preset ratio can be 0.1 or other values.

[0061] In this way, after calculating the popularity value of the first video, if the popularity value of the first video is greater than that of other videos, the first video will continue to be processed based on the positive comment data of the derivative video clips, thereby shortening the development cycle of derivative content. In addition, points can be calculated based on the popularity value to give to viewers, so as to improve user activity, realize a feedback loop, and ultimately help improve the user viewing experience and the revenue of film and television dramas.

[0062] In another embodiment of this application, the implementation method of S120 will be explained in detail.

[0063] Figure 2 A flowchart illustrating another video processing method provided in an embodiment of this disclosure is shown.

[0064] like Figure 2 As shown, the video processing method may include the following steps.

[0065] S210. Obtain the derivative plot text corresponding to at least one first playback position in the first video.

[0066] S210 is similar to S110, and will not be described in detail here.

[0067] S220. Obtain the first event and the first role associated with the first playback position from the node information contained in the preset knowledge graph, and obtain the first role description corresponding to the first role from the attribute information contained in the preset knowledge graph.

[0068] The preset knowledge graph is determined in advance according to the plot logic of the first video.

[0069] Optionally, node information may include roles and events. A role can be understood as a character in the first video, including the main character, supporting characters, etc. The event to which a role belongs is related to that role.

[0070] Optionally, the attribute information includes character attribute information and event attribute information. Character attribute information specifically includes a character description, and event attribute information specifically includes an event description. A character description can be understood as a character's personality. For example, a reserved personality, an outgoing personality, a suspicious personality, or a ruthless personality. An event description includes event-related information. For example, the location of the event, the scene in which it occurred, and the specific event itself.

[0071] In some embodiments, the preset knowledge graph is determined through the following steps: obtaining video information and script information corresponding to the first video; based on the video information and script information, determining the characters, relationships between characters, character descriptions, events, relationships between events, event descriptions, and event-role relationships in the first video; generating the preset knowledge graph by using characters as first-type points, relationships between characters as edges between two first-type points, events as second-type points, relationships between events as edges between two second-type points, character descriptions as attributes representing first-type points of characters, event descriptions as attributes representing second-type points of events, and event-role relationships as the association between characters represented by first-type points and events represented by second-type points.

[0072] Optionally, the character description includes one or more of the following combinations: personality traits, character positioning, and character nickname; the event description includes one or more of the following combinations: the location of the event, the scene of the event, and the time of the event; the relationship between characters includes one or more of the following combinations: antagonistic relationship and cooperative relationship; the relationship between events includes one or more of the following combinations: causal relationship and parallel relationship.

[0073] In this way, based on the character information and event information extracted from the video information and script information of the first video, a pre-defined knowledge graph of the first video is constructed, so that the character information and event information of the first video can be abstractly represented by the knowledge graph, which is convenient for subsequent analysis.

[0074] In some embodiments, the step S220, "obtaining the first character description corresponding to the first character from the attribute information contained in the preset knowledge graph," is specifically implemented through the following steps: taking the first event as the starting point, and following the reverse direction of the plot logic, based on the event-to-event relationships, character-to-character relationships, and event-to-character relationships contained in the preset knowledge graph, obtaining a first preset number of second events before the first playback position from the preset knowledge graph, wherein the second events are associated with the first character; and determining the first character description corresponding to the first character based on the second character description associated with the first character in the first preset number of second events.

[0075] The first preset number of second events consists of historical event information prior to the first playback position, which can be determined in advance according to requirements. For example, the second events prior to the first playback position could be 2-3 events prior to the first playback position.

[0076] In this way, by combining historical event information prior to the first playback position, the character description can be accurately determined from the preset knowledge graph.

[0077] S230. The first event is determined as the event information associated with the first playback position, and the first role and the first role description are determined as the role information associated with the first playback position.

[0078] Understandably, since the first event, the first role, and the description of the first role are determined based on historical events and the roles in those historical events, the event information and role information associated with the first playback position obtained thereby have high accuracy.

[0079] S240. Based on event information, character information, and derivative plot text, generate derivative guidance information corresponding to the first playback position.

[0080] S240 is similar to S130, so it will not be described in detail here.

[0081] S250: Based on the plot logic, derivative guidance information, reference segments at the first playback position, and derivative plot text, generate derivative video segments corresponding to the first playback position.

[0082] The derived video clips are not part of the original video content of the first video. Reference clips include the first video content within a first time period before the first playback position and / or the second video content within a second time period after the first playback position.

[0083] In another embodiment of this application, the implementation method of S140 will be explained in detail.

[0084] Figure 3 A flowchart illustrating another video processing method provided in an embodiment of this disclosure is shown.

[0085] like Figure 3 As shown, the video processing method may include the following steps.

[0086] S310. Obtain the derivative plot text corresponding to at least one first playback position in the first video.

[0087] S320. Obtain event information and character information associated with the first playback position from the preset knowledge graph corresponding to the first video, wherein the preset knowledge graph is predetermined according to the plot logic of the first video.

[0088] S330. Based on event information, character information, and derivative plot text, generate derivative guidance information corresponding to the first playback position.

[0089] S310~S330 are similar to S110~S130, and will not be described in detail here.

[0090] S340. Based on the plot logic, derived guiding information, reference clips, and derived plot text, determine the predicted derived video corresponding to the first playback position.

[0091] Among them, the predicted derivative video is the derivative content obtained from the preliminary prediction of the first playback position.

[0092] Optionally, the number of predicted derivative videos is at least one.

[0093] In some embodiments, S340 is specifically implemented through the following steps: S3401, inputting the plot logic, derived guidance information, reference segments, and derived plot text into a video prediction network for prediction, generating at least two prediction results corresponding to the first playback position; S3402, inputting the at least two prediction results into a plot analysis network for plot causal analysis and plot style analysis, determining the plot analysis values ​​corresponding to the at least two prediction results respectively; S3403, obtaining the prediction result with a plot analysis value greater than a preset threshold from the at least two prediction results, and using it as the predicted derived video.

[0094] Specifically, S3401 is implemented through the following steps: S1, inputting the reference segment into the style analysis subnetwork of the video prediction network to extract the plot style and determine the plot style information of the reference segment; S2, inputting the plot style information, derived guiding information, plot logic, and derived plot text into the temporal analysis subnetwork of the video prediction network to perform temporal analysis and determine at least two predicted events and the time information corresponding to each of the at least two predicted events; S3, inputting the at least two predicted events and the time information into the coherence analysis subnetwork of the video prediction network to perform event sequence analysis and generate at least two prediction results that meet the preset sequence conditions.

[0095] Optionally, the style analysis subnetwork includes, but is not limited to, video style understanding networks (such as the GPT-4V network).

[0096] The narrative style information refers to the visual characteristics of the reference segment. Optionally, the narrative style information includes, but is not limited to, the scene characteristics, character emotional characteristics, camera style characteristics, and narrative style characteristics of the reference segment.

[0097] Among them, at least two predicted events and their respective time information are features containing temporal order.

[0098] Among them, at least two prediction results that meet the preset order conditions are prediction derivatives that satisfy grammatical and logical rationality.

[0099] Specifically, for S1, the electronic device acquires cue words (e.g., extracting information such as "scene features, character emotion features, shot style features, plot style features"), and inputs the cue words and reference segments into the style analysis subnetwork in the video prediction network to extract the plot style information of the reference segments.

[0100] Specifically, for S2, the electronic device first concatenates the plot style information, derivative guidance information, plot logic, and derivative plot text into a unified text sequence. Then, it performs word segmentation and vectorization on the unified text sequence to obtain numerical vectors that can be recognized by the temporal analysis subnetwork. Then, it uses the logic analysis subnetwork to process the recognizable numerical vectors according to temporal logic to obtain features containing temporal order, which serve as at least two predicted events and their corresponding time information.

[0101] Optionally, the logic analysis subnetwork includes, but is not limited to, a Long Short-Term Memory (LSTM) network.

[0102] Specifically, for S3, the electronic device uses a coherence analysis subnetwork to convert at least two predicted events and their corresponding time information into sentence-level label sequences. Then, based on a predefined legal transition matrix, it corrects illegal transitions in the sentence-level label sequences to complete the sentence sequence conversion and logical rationality transition, generating at least two predicted videos.

[0103] For example, if character A moves from being injured to running, it's an illegal transfer. It should be changed to character A being injured and then being helped away.

[0104] Optionally, the coherence analysis subnetwork includes, but is not limited to, a Conditional Random Field (CRF) network.

[0105] In this way, based on the style analysis subnetwork, temporal analysis subnetwork, and coherence analysis subnetwork in the video prediction network, style extraction, temporal analysis, and coherence analysis are performed on the input information to obtain preliminary predicted derivative content that conforms to the plot logic, thereby improving the rationality of the predicted derivative content.

[0106] Specifically, S3402 is implemented through the following steps: S4, inputting at least two prediction results into the causal analysis subnetwork of the plot analysis network to perform plot causal analysis, and determining the plot causal analysis values ​​corresponding to the at least two prediction results respectively; S5, inputting at least two prediction results into the style matching degree subnetwork of the plot analysis network to perform plot style analysis, and determining the plot style analysis values ​​corresponding to the at least two prediction results respectively; S6, performing a weighted sum of the plot causal analysis values ​​and plot style analysis values ​​corresponding to the at least two prediction results respectively, to obtain the plot analysis values ​​corresponding to the at least two prediction results respectively.

[0107] Optionally, the narrative analysis network includes, but is not limited to, pre-trained BERT (e.g., RoBERTa).

[0108] Among them, the plot causal analysis value is the logical consistency score of the prediction results, which is used to evaluate the degree of matching between the event logic and character personality in the prediction results and the character information and event information in the preset knowledge graph.

[0109] Optionally, the causal analysis score can be 0 to 0.5.

[0110] Among them, the plot style analysis value is the style matching degree of the test results, which is used to evaluate the degree of matching between the prediction results and the visual features integrated in the plot style information.

[0111] Optionally, the plot style analysis score can be 0 to 0.5.

[0112] The plot analysis value is a composite score of the logical consistency score and style matching degree of the prediction results.

[0113] Optionally, the plot analysis score can be 0 to 1 point.

[0114] Specifically, for S4, the electronic device inputs at least two prediction results into the causal analysis subnetwork in the plot analysis network. Based on the causal analysis subnetwork, the event logic and character personality contained in the at least two prediction results are matched with the character information and event information in the preset knowledge graph to determine the logical consistency score, which is used as the causal analysis value.

[0115] Specifically, for S5, the electronic device inputs at least two prediction results into the style matching degree subnetwork in the plot analysis network, so as to determine whether the plot style information corresponding to the at least two prediction results fits the plot style of the first video based on the style matching degree subnetwork, thereby obtaining the style matching degree, and using the style matching degree as the plot style analysis value.

[0116] Specifically, for S6, for each prediction result, the electronic device obtains the first weight corresponding to the plot causal analysis value and the second weight corresponding to the plot style analysis value, multiplies the plot causal analysis value by the first weight to obtain the first product, and multiplies the plot style analysis value by the second weight to obtain the second product. Then, the first product and the second product are added together to obtain the plot analysis value corresponding to each prediction result.

[0117] Specifically, S3403 is implemented through the following steps: comparing the plot analysis values ​​corresponding to at least two prediction results with preset thresholds, and outputting the prediction results whose plot analysis values ​​are greater than the preset thresholds to obtain the predicted derivative video of the first video.

[0118] The preset threshold is the minimum plot analysis value used to filter the final derivative content.

[0119] Optionally, the preset threshold is 0.8 or 0.7, or other values.

[0120] For example, the plot analysis value corresponding to prediction result A (character x waiting to retaliate) is 0.85, the plot analysis value corresponding to prediction result B (character x directly retaliates) is 0.72, and the plot analysis value corresponding to prediction result C (character x impulsively seeks revenge) is 0.6. The preset threshold is 0.8. The plot analysis values ​​corresponding to prediction results A, B, and C are compared with the preset threshold of 0.8 to determine the prediction result. If the plot analysis value corresponding to prediction result A is greater than the preset threshold, then prediction result A is used as the predicted derivative video.

[0121] In this way, for at least two preliminary predicted derivative content generated by the video prediction network, the video processing device continues to use the plot analysis network to perform plot analysis on each preliminary predicted derivative content, and selects the derivative content with higher plot analysis value as the predicted derivative video, thus improving the plot rationality of the predicted derivative video.

[0122] S350. Associate the predicted derivative video with the relevant information corresponding to the first playback position, and determine the predicted derivative video with the associated relevant information as the derivative video segment corresponding to the first playback position. The relevant information is determined based on the derivative plot text.

[0123] The relevant information section is a brief introduction or summary of the predicted derivative video, which serves to introduce the main content of the predicted derivative video.

[0124] Specifically, the video processing equipment can perform semantic understanding on the derived plot text, extract keywords from the semantic understanding information, and finally fill the keywords into a preset template to generate relevant information for predicting the derived video.

[0125] Optionally, the relevant information can be represented as images or text.

[0126] In this way, after determining the predicted derivative video, relevant information such as a brief introduction or summary is added to the predicted derivative video. The predicted derivative video with the relevant information is then used as the derivative video segment at the first playback position. This allows users to quickly understand the main content of the derivative video segment by displaying the brief introduction or summary of the derivative video segment when the first video reaches the first playback position, making it convenient for users to quickly learn about the derivative content they are interested in.

[0127] In yet another embodiment of this application, a video playback method is provided. The following describes... Figure 4 The video playback method provided in this disclosure embodiment will be described. In this disclosure embodiment, the video playback method can be executed by a video playback device. The video playback device can be an electronic device or a server.

[0128] Figure 4 A flowchart illustrating a video processing method provided in an embodiment of this disclosure is shown.

[0129] like Figure 4 As shown, the video playback method may include the following steps.

[0130] S410. When the first video is played to the first playback position, display relevant information about at least one derivative video segment associated with the first playback position.

[0131] The first video includes at least one first playback position. The derivative video segments do not belong to the original video content of the first video. The derivative video segments are determined based on the plot logic of the first video, the derivative guidance information corresponding to the first playback position, the reference segments of the first playback position, and the derivative plot text corresponding to the first playback position. The derivative guidance information is generated based on the derivative plot text, the event information associated with the first playback position, and the character information. The event information and character information are obtained from the preset knowledge graph corresponding to the first video. The preset knowledge graph is predetermined according to the plot logic of the first video. The reference segments include the first video content of the first time period before the first playback position and / or the second video content of the second time period after the first playback position.

[0132] The relevant information section is a brief introduction or summary of the predicted derivative video, which serves to introduce the main content of the predicted derivative video.

[0133] Optionally, the relevant information can be represented as images or text.

[0134] In some embodiments, the event information and role information associated with the first playback position are specifically implemented through the following steps: obtaining the first event and the first role associated with the first playback position from the node information contained in the preset knowledge graph, and obtaining the first role description corresponding to the first role from the attribute information contained in the preset knowledge graph; determining the first event as the event information associated with the first playback position, and determining the first role and the first role description as the role information associated with the first playback position.

[0135] In some embodiments, the first event and the first character associated with the first playback position are obtained from the node information contained in the preset knowledge graph. This is specifically achieved through the following steps: taking the first event as the starting point, and following the reverse direction of the plot logic, based on the inter-event relationships, inter-character relationships, and event-character relationships contained in the preset knowledge graph, a first preset number of second events prior to the first playback position are obtained from the preset knowledge graph, wherein the second events are associated with the first character; based on the second character description associated with the first character in the first preset number of second events, the first character description corresponding to the first character is determined.

[0136] In some embodiments, the derivative guidance information corresponding to the first playback position is determined through the following steps: determining the derivative plot type corresponding to the derivative plot text, wherein the derivative plot type is behavioral or result-based; and concatenating the event information, character information, and derivative plot type according to a preset splicing format to generate derivative guidance information.

[0137] In some embodiments, the derivative video segment corresponding to the first playback position is specifically implemented through the following steps: based on the plot logic, derivative guidance information, reference segments, and derivative plot text, the predicted derivative video corresponding to the first playback position is determined; the predicted derivative video is associated with the relevant information corresponding to the first playback position, and the predicted derivative video associated with the relevant information is determined as the derivative video segment corresponding to the first playback position, wherein the relevant information is determined according to the derivative plot text.

[0138] In some embodiments, based on plot logic, derived guidance information, reference segments, and derived plot text, the predicted derived video corresponding to the first playback position is determined. This is achieved through the following steps: inputting the plot logic, derived guidance information, reference segments, and derived plot text into a video prediction network for prediction, generating at least two prediction results corresponding to the first playback position; inputting the at least two prediction results into a plot analysis network for plot causality analysis and plot style analysis, determining the plot analysis values ​​corresponding to the at least two prediction results respectively; and obtaining the prediction result with a plot analysis value greater than a preset threshold from the at least two prediction results as the predicted derived video.

[0139] In some embodiments, the plot logic, derived guidance information, reference segments, and derived plot text are input into a video prediction network for prediction to generate at least two prediction results corresponding to the first playback position. This is specifically achieved through the following steps: the reference segment is input into the style analysis subnetwork of the video prediction network for plot style extraction to determine the plot style information of the reference segment; the plot style information, derived guidance information, plot logic, and derived plot text are input into the temporal analysis subnetwork of the video prediction network for temporal analysis to determine at least two prediction events and the time information corresponding to each of the at least two prediction events; the at least two prediction events and the time information are input into the coherence analysis subnetwork of the video prediction network for event sequence analysis to generate at least two prediction results that satisfy preset order conditions.

[0140] In some embodiments, at least two prediction results are input into a plot analysis network for plot causality analysis and plot style analysis to determine the plot analysis values ​​corresponding to the at least two prediction results. This is specifically achieved through the following steps: inputting at least two prediction results into a causal analysis subnetwork of the plot analysis network for plot causality analysis to determine the plot causality analysis values ​​corresponding to the at least two prediction results; inputting at least two prediction results into a style matching degree subnetwork of the plot analysis network for plot style analysis to determine the plot style analysis values ​​corresponding to the at least two prediction results; and performing a weighted sum of the plot causality analysis values ​​and plot style analysis values ​​corresponding to the at least two prediction results to obtain the plot analysis values ​​corresponding to the at least two prediction results.

[0141] In some embodiments, the preset knowledge graph of the first video is specifically implemented through the following steps: obtaining video information and script information corresponding to the first video; based on the video information and script information, determining the characters, relationships between characters, character descriptions, events, relationships between events, event descriptions, and event-role relationships in the first video; generating a preset knowledge graph by using characters as first-type points, relationships between characters as edges between two first-type points, events as second-type points, relationships between events as edges between two second-type points, character descriptions as attributes representing first-type points of characters, event descriptions as attributes representing second-type points of events, and event-role relationships as the association between characters represented by first-type points and events represented by second-type points.

[0142] In some embodiments, the character description includes one or more of the following combinations: personality traits, character positioning, and character nickname; the event description includes one or more of the following combinations: event location, event scene, and event time; the relationship between characters includes one or more of the following combinations: antagonistic relationship and cooperative relationship; the relationship between events includes one or more of the following combinations: causal relationship and parallel relationship.

[0143] S420, in response to the selection operation of a first relevant information among at least one relevant information, play the first derived video segment corresponding to the first relevant information.

[0144] Understandably, each piece of relevant information is associated with a derivative video clip. When a user selects the first piece of relevant information on the video playback page of the video playback device, the video playback device responds to this selection by playing the first derivative video clip corresponding to the first piece of relevant information from the first playback position.

[0145] In this way, since the derivative video clips are generated according to the plot logic of the first video, and the relevant information associated with the derivative video clips can help users quickly understand the main content of the derivative video clips, when users want to watch derivative video clips that they are interested in, they can select relevant information to retrieve the derivative video clip corresponding to that information and play it in the first playback position, thus improving the user's video viewing experience.

[0146] In some embodiments of this application, after executing S420, the method further includes: when the first derived video segment is played to the second playback position, continuing to play the unplayed content in the first video, wherein the second playback position is the end position of the first derived video segment.

[0147] Unplayed content refers to the video content after the first playback position in the first video.

[0148] For example, if the first video is "XX" and the first playback position is the 20-minute mark of episode 10 of "XX", when the first derivative video clip plays to the second playback position, the unplayed content in the first video is the video content after the 20-minute mark of episode 10 of "XX". Then the video playback device will start from the second playback position and continue playing the unplayed content of the first video.

[0149] In this way, after the derivative video clip finishes playing, the unplayed content of the first video continues to play from the end of the derivative video clip, ensuring that the video content of the first video is not interrupted by the playback of the derivative video clip and guaranteeing the integrity and continuity of the video playback.

[0150] In some embodiments of this application, after performing S420, the method further includes: when the first derived video segment is played to a second playback position, displaying second related information of the second derived video segment in at least one derived video segment, wherein the second playback position is the end position of the first derived video segment; and playing the second derived video segment in response to a selection operation of the second related information.

[0151] The second derivative video segment is any derivative video segment other than the first derivative video segment.

[0152] Understandably, in order to facilitate users to watch more derivative content, after the video playback device finishes playing the first derivative video segment, it can also display at least one second related information corresponding to the second derivative video segment. If the user selects a second related information, the second derivative video segment corresponding to that second related information is retrieved, and the second derivative video segment continues to play from the end position of the first derivative video segment.

[0153] This allows for the sequential playback of multiple derivative video clips, providing a richer and more immersive interactive experience.

[0154] In another embodiment of this application, the overall logic of the video processing method and the video playback method is explained in detail.

[0155] Figure 5 A logical schematic diagram of a video processing method and a video playback method provided in an embodiment of this disclosure is shown.

[0156] like Figure 5 As shown, the video processing method may include the following steps.

[0157] S510, Data Preprocessing.

[0158] In this embodiment, the video processing device acquires the main feature and script of the first video, and extracts video frames from the main feature of the target film and television drama to obtain the video information and script information of the first video.

[0159] S520, Construct a pre-defined knowledge graph.

[0160] In this embodiment, the video processing device extracts plot logic from video information and script information, determines the plot logic as global constraint information, and constructs a preset knowledge graph based on the global constraint information.

[0161] S530, Obtain derivative story text.

[0162] S540, Generate real-time constraint information.

[0163] In this embodiment, the video processing device first obtains the event information and character information associated with the first playback position from the preset knowledge graph, and then generates the derivative guidance information corresponding to the first playback position based on the event information, character information and derivative plot text, and uses the derivative guidance information as real-time constraint information.

[0164] S550, predicting derived video clips.

[0165] In this embodiment, the video processing device combines real-time constraint information, global constraint information corresponding to the first video, reference segments at the first playback position, and derived plot text to predict the derived video segments corresponding to the first playback position.

[0166] Among them, the derivative video clips are not part of the original video content of the first video, and the derivative video clips are associated with relevant information.

[0167] The reference segments include the first video content of the first time period before the first playback position and / or the second video content of the second time period after the first playback position.

[0168] S560 pushes and plays derivative video clips.

[0169] In this embodiment, the video processing device pushes the derived video clip to the video playback device, and the video playback device displays the relevant information of the derived video clip, and displays the derived video clip corresponding to the relevant information according to the user's selection operation of the relevant information.

[0170] S570, Implement interactive feedback.

[0171] Specifically, the video playback device acquires user voting data, viewing completion rate, and positive comment data for derivative video clips, and sends these data to the video processing device. The video processing device calculates the popularity value of the derivative video clip based on the voting data, viewing completion rate, and positive comment data. If the popularity value of the derivative video clip is greater than the popularity value of other derivative content, the first video is processed based on the derivative video clip and its positive comment data.

[0172] S580, Execution Points Rewards.

[0173] Specifically, the video processing device determines the points to be awarded to the viewing user based on the popularity value of the derived video clip, and sends the points to the video playback device. The video playback device then displays the user's points based on the user's point acquisition operation.

[0174] This disclosure also provides a video playback device for implementing the above-described video processing method, which will be described below in conjunction with... Figure 6 The following explanation is provided. In this embodiment of the disclosure, the video processing apparatus can be executed by a video processing device. The video processing device can be an electronic device or a server.

[0175] Figure 6 A schematic diagram of the structure of a video processing apparatus provided in an embodiment of this disclosure is shown.

[0176] like Figure 6 As shown, the video processing apparatus 600 may include: The first acquisition module 610 is used to acquire at least one derivative plot text corresponding to a first playback position in the first video; The second acquisition module 620 is used to acquire event information and character information associated with the first playback position from the preset knowledge graph corresponding to the first video, wherein the preset knowledge graph is predetermined according to the plot logic of the first video; The first generation module 630 is used to generate derivative guidance information corresponding to the first playback position based on the event information, the character information, and the derivative plot text. The second generation module 640 is used to generate a derivative video segment corresponding to the first playback position based on the plot logic, the derivative guidance information, the reference segment of the first playback position, and the derivative plot text. The derivative video segment does not belong to the original video content of the first video. The reference segment includes the first video content of the first time period before the first playback position and / or the second video content of the second time period after the first playback position.

[0177] An embodiment of this disclosure provides a video processing apparatus that acquires derived plot text corresponding to at least one first playback position in a first video; acquires event information and character information associated with the first playback position from a preset knowledge graph corresponding to the first video, the preset knowledge graph being predetermined according to the plot logic of the first video; generates derived guidance information corresponding to the first playback position based on the event information, character information, and derived plot text; and generates a derived video segment corresponding to the first playback position based on the plot logic, derived guidance information, reference segments of the first playback position, and derived plot text. The derived video segment is not part of the original video content of the first video, and the reference segments include the first video content of a first time period before the first playback position and / or the second video content of a second time period after the first playback position. Thus, by first determining event information and character information based on a preset knowledge graph representing the plot logic, then generating guidance text for the derived video segment based on the event information, character information, and derived plot text, and combining the guidance text, plot logic, reference materials, and derived plot text, derived content conforming to the plot logic is generated, improving the production quality and efficiency of derived content.

[0178] In some embodiments of this disclosure, the second acquisition module 620 includes: The first acquisition unit is used to acquire the first event and the first role associated with the first playback position from the node information contained in the preset knowledge graph, and to acquire the first role description corresponding to the first role from the attribute information contained in the preset knowledge graph. The first determining unit is configured to determine the first event as event information associated with the first playback position, and to determine the first role and the first role description as role information associated with the first playback position.

[0179] In some embodiments of this disclosure, the first acquisition unit is specifically used for: Taking the first event as the starting point, and following the reverse direction of the plot logic, based on the inter-event relationships, inter-character relationships, and event-character relationships contained in the preset knowledge graph, a first preset number of second events before the first playback position are obtained from the preset knowledge graph, wherein the second events are associated with the first character; Based on the second role description associated with the first role in the first preset number of second events, determine the first role description corresponding to the first role.

[0180] In some embodiments of this disclosure, the first generation module 630 is specifically used for: Determine the derivative plot type corresponding to the derivative plot text, wherein the derivative plot type is either behavior-based or result-based; According to the preset splicing format, the event information, the character information, and the derivative plot type are spliced ​​together to generate the derivative guidance information.

[0181] In some embodiments of this disclosure, the second generation module 640 includes: The second determining unit is used to determine the predicted derivative video corresponding to the first playback position based on the plot logic, the derivative guidance information, the reference segment, and the derivative plot text. The third determining unit is used to associate the predicted derivative video with the relevant information corresponding to the first playback position, and to determine the predicted derivative video associated with the relevant information as the derivative video segment corresponding to the first playback position, wherein the relevant information is determined according to the derivative plot text.

[0182] In some embodiments of this disclosure, the second determining unit includes: The first generation subunit is used to input the plot logic, the derivative guidance information, the reference segment and the derivative plot text into the video prediction network for prediction, and generate at least two prediction results corresponding to the first playback position. The first determining subunit is used to input the at least two prediction results into the plot analysis network to perform plot causal analysis and plot style analysis, and to determine the plot analysis values ​​corresponding to the at least two prediction results respectively. The second determining subunit is used to obtain the prediction result from the at least two prediction results where the plot analysis value is greater than a preset threshold, and use it as the prediction-derived video.

[0183] In some embodiments of this disclosure, the first generation subunit is specifically used for: The reference segment is input into the style analysis subnetwork in the video prediction network to extract the plot style and determine the plot style information of the reference segment. The plot style information, the derivative guidance information, the plot logic, and the derivative plot text are input into the temporal analysis subnetwork in the video prediction network for temporal analysis to determine at least two predicted events and the time information corresponding to the at least two predicted events respectively. The at least two predicted events and the time information are input into the coherence analysis subnetwork in the video prediction network for event sequence analysis, generating at least two prediction results that satisfy preset sequence conditions.

[0184] In some embodiments of this disclosure, the first determining subunit is specifically used for: The at least two prediction results are input into the causal analysis subnetwork in the plot analysis network to perform plot causal analysis, and the plot causal analysis values ​​corresponding to the at least two prediction results are determined respectively. The at least two prediction results are input into the style matching degree subnetwork in the plot analysis network to perform plot style analysis, and the plot style analysis values ​​corresponding to the at least two prediction results are determined respectively. The plot causal analysis value and the plot style analysis value corresponding to the at least two prediction results are weighted and summed to obtain the plot analysis value corresponding to the at least two prediction results.

[0185] In some embodiments of this disclosure, the device further includes: The third acquisition module is used to acquire video information and script information corresponding to the first video; The first determining module is used to determine, based on the video information and the script information, the characters, relationships between characters, character descriptions, events, relationships between events, event descriptions, and relationships between events and characters in the first video; The third generation module is used to generate the preset knowledge graph by using the role as a first type of point, the relationship between the roles as an edge between two first type of points, the event as a second type of point, the relationship between the events as an edge between two second type of points, the role description as an attribute representing the first type of point of the role, and the event description as an attribute representing the second type of point of the event, and the event-role relationship as the association between the role represented by the first type of point and the event represented by the second type of point.

[0186] In some embodiments of this disclosure, the character description includes one or more combinations of the following: personality traits, character positioning, and character nickname; The event description includes one or more combinations of the following: the location of the event, the scene of the event, and the time of the event; The relationships between the roles include one or more combinations of the following: adversarial relationship, cooperative relationship; The relationships between events include one or more of the following combinations: causal relationship, parallel relationship.

[0187] It should be noted that, Figure 6 The video processing device 600 shown can perform... Figures 1-3The various steps in the method embodiment shown are implemented. Figures 1-3 The processes and effects in the method embodiments shown are not described in detail here.

[0188] This disclosure also provides a video playback device for implementing the above-described video playback method, which will be described below in conjunction with... Figure 7 The following explanation is provided. In this embodiment of the disclosure, the video playback device can be executed by a video playback device. The video playback device can be an electronic device or a server.

[0189] Figure 7 A schematic diagram of the structure of a video playback device provided in an embodiment of this disclosure is shown.

[0190] like Figure 7 As shown, the video playback device 700 may include: Display module 710 is configured to display relevant information of at least one derivative video segment associated with the first playback position when the first video is played to the first playback position. The first video includes at least one first playback position. The derivative video segments do not belong to the original video content of the first video. The derivative video segments are determined based on the plot logic of the first video, the derivative guidance information corresponding to the first playback position, the reference segment of the first playback position, and the derivative plot text corresponding to the first playback position. The derivative guidance information is generated based on the derivative plot text, the event information associated with the first playback position, and the character information. The event information and the character information are obtained from a preset knowledge graph corresponding to the first video. The preset knowledge graph is predetermined according to the plot logic of the first video. The reference segment includes the first video content of a first time period before the first playback position and / or the second video content of a second time period after the first playback position. The first playback module 720 is configured to play a first derived video segment corresponding to the first relevant information in response to a selection operation of at least one of the relevant information.

[0191] In this way, since the derivative video clips are generated according to the plot logic of the first video, and the relevant information associated with the derivative video clips can help users quickly understand the main content of the derivative video clips, when users want to watch derivative video clips that they are interested in, they can select relevant information to retrieve the derivative video clip corresponding to that information and play it in the first playback position, thus improving the user's video viewing experience.

[0192] In some embodiments of this disclosure, the device further includes: The second playback module is used to continue playing unplayed content in the first video when the first derived video segment reaches the second playback position, wherein the second playback position is the end position of the first derived video segment.

[0193] In some embodiments of this disclosure, the device further includes: The display module is configured to display second relevant information of the second derivative video segment in the at least one derivative video segment when the first derivative video segment is played to the second playback position, wherein the second playback position is the end position of the first derivative video segment; The third playback module is used to play the second derived video segment in response to the selection operation of the second relevant information.

[0194] It should be noted that, Figure 7 The video processing device 700 shown can perform... Figure 4 The various steps in the method embodiment shown are implemented. Figure 4 The processes and effects in the method embodiments shown are not described in detail here.

[0195] Figure 8 A schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure is shown.

[0196] like Figure 8 As shown, the electronic device may include a processor 801 and a memory 802 storing computer program instructions.

[0197] Specifically, the processor 801 may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0198] Memory 802 may include a large-capacity storage device for advertising or instructions. For example, and not limitingly, memory 802 may include a hard disk drive (HDD), a floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or a Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, memory 802 may include removable or non-removable (or fixed) media. Where appropriate, memory 802 may be internal or external to the integrated gateway device. In a particular embodiment, memory 802 is a non-volatile solid-state memory. In a particular embodiment, memory 802 includes read-only memory (ROM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (Electrically Programmable ROM, EPROM), an electrically erasable programmable PROM (EEPROM), an electrically alterable ROM (EAROM), or flash memory, or a combination of two or more of these.

[0199] The processor 801 acquires and executes computer program instructions stored in the memory 802 to perform the steps of the video processing method provided in the embodiments of this disclosure.

[0200] In one example, the electronic device may also include a transceiver 803 and a bus 804. Wherein, as... Figure 8 As shown, the processor 801, memory 802 and transceiver 803 are connected via bus 804 and communicate with each other.

[0201] Bus 804 includes hardware, software, or both. For example, and not limitingly, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industrial Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a MicroChannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local Bus (VLB) bus, or other suitable buses, or a combination of two or more of these. Where appropriate, bus 804 may include one or more buses. Although specific buses are described and illustrated in the embodiments of this application, this application considers any suitable bus or interconnection.

[0202] The following are embodiments of a computer-readable storage medium provided in this disclosure. This computer-readable storage medium belongs to the same inventive concept as the video processing methods in the above embodiments. For details not described in detail in the embodiments of the computer-readable storage medium, please refer to the embodiments of the above video processing methods.

[0203] This embodiment provides a storage medium containing computer-executable instructions, which, when executed by a computer processor, are used to perform a video processing method, including: Obtain the derived plot text corresponding to at least one first playback position in the first video; The event information and character information associated with the first playback position are obtained from the preset knowledge graph corresponding to the first video, wherein the preset knowledge graph is predetermined according to the plot logic of the first video; Based on the event information, the character information, and the derivative plot text, generate derivative guidance information corresponding to the first playback position; Based on the plot logic, the derived guidance information, the reference segment at the first playback position, and the derived plot text, a derived video segment corresponding to the first playback position is generated. The derived video segment does not belong to the original video content of the first video. The reference segment includes the first video content of a first time period before the first playback position and / or the second video content of a second time period after the first playback position.

[0204] This embodiment provides a storage medium containing computer-executable instructions, which, when executed by a computer processor, are used to perform a video playback method, including: When the first video is played to the first playback position, information related to at least one derivative video segment associated with the first playback position is displayed. The first video includes at least one first playback position. The derivative video segments do not belong to the original video content of the first video. The derivative video segments are determined based on the plot logic of the first video, the derivative guidance information corresponding to the first playback position, the reference segment of the first playback position, and the derivative plot text corresponding to the first playback position. The derivative guidance information is generated based on the derivative plot text, the event information associated with the first playback position, and the character information. The event information and the character information are obtained from a preset knowledge graph corresponding to the first video. The preset knowledge graph is predetermined according to the plot logic of the first video. The reference segment includes the first video content of a first time period before the first playback position and / or the second video content of a second time period after the first playback position. In response to a selection operation of a first relevant information among at least one of the relevant information, a first derived video segment corresponding to the first relevant information is played.

[0205] Of course, the computer-executable instructions provided in the embodiments of this disclosure are not limited to the above-described method operations, but can also perform related operations in the video processing method or video playback method provided in any embodiment of this disclosure.

[0206] Based on the above description of the implementation methods, those skilled in the art can clearly understand that this disclosure can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this disclosure, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer cloud platform (which may be a personal computer, server, or network cloud platform, etc.) to execute the video processing method or video playback method provided in the various embodiments of this disclosure.

[0207] Note that the above description is merely a preferred embodiment and the technical principles employed in this disclosure. Those skilled in the art will understand that this disclosure is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions can be made without departing from the scope of protection of this disclosure. Therefore, although this disclosure has been described in detail through the above embodiments, it is not limited to the above embodiments. Many other equivalent embodiments may be included without departing from the concept of this disclosure, and the scope of this disclosure is determined by the scope of the appended claims.

Claims

1. A video processing method, characterized in that, include: Obtain the derived plot text corresponding to at least one first playback position in the first video; The event information and character information associated with the first playback position are obtained from the preset knowledge graph corresponding to the first video, wherein the preset knowledge graph is predetermined according to the plot logic of the first video; Based on the event information, the character information, and the derivative plot text, generate derivative guidance information corresponding to the first playback position; Based on the plot logic, the derived guidance information, the reference segment at the first playback position, and the derived plot text, a derived video segment corresponding to the first playback position is generated. The derived video segment does not belong to the original video content of the first video. The reference segment includes the first video content of a first time period before the first playback position and / or the second video content of a second time period after the first playback position.

2. The method according to claim 1, characterized in that, The step of obtaining event information and role information associated with the first playback position from the preset knowledge graph corresponding to the first video includes: The first event and the first role associated with the first playback position are obtained from the node information contained in the preset knowledge graph, and the first role description corresponding to the first role is obtained from the attribute information contained in the preset knowledge graph. The first event is identified as event information associated with the first playback location, and the first role and its description are identified as role information associated with the first playback location.

3. The method according to claim 2, characterized in that, The step of obtaining the first role description corresponding to the first role from the attribute information contained in the preset knowledge graph includes: Taking the first event as the starting point, and following the reverse direction of the plot logic, based on the inter-event relationships, inter-character relationships, and event-character relationships contained in the preset knowledge graph, a first preset number of second events before the first playback position are obtained from the preset knowledge graph, wherein the second events are associated with the first character; Based on the second role description associated with the first role in the first preset number of second events, determine the first role description corresponding to the first role.

4. The method according to claim 1, characterized in that, The step of generating derivative guidance information corresponding to the first playback position based on the event information, the character information, and the derivative plot text includes: Determine the derivative plot type corresponding to the derivative plot text, wherein the derivative plot type is either behavior-based or result-based; According to the preset splicing format, the event information, the character information, and the derivative plot type are spliced ​​together to generate the derivative guidance information.

5. The method according to claim 1, characterized in that, The step of generating a derivative video segment corresponding to the first playback position based on the plot logic, the derivative guidance information, the reference segment at the first playback position, and the derivative plot text includes: Based on the plot logic, the derivative guidance information, the reference clips, and the derivative plot text, determine the predicted derivative video corresponding to the first playback position; The predicted derivative video is associated with the relevant information corresponding to the first playback position, and the predicted derivative video associated with the relevant information is determined as the derivative video segment corresponding to the first playback position. The relevant information is determined based on the derivative plot text.

6. The method according to claim 5, characterized in that, The step of determining the predicted derivative video corresponding to the first playback position based on the plot logic, the derivative guidance information, the reference clips, and the derivative plot text includes: The plot logic, the derivative guidance information, the reference segment, and the derivative plot text are input into the video prediction network for prediction, generating at least two prediction results corresponding to the first playback position. The at least two prediction results are input into the plot analysis network to perform plot causality analysis and plot style analysis, and the plot analysis values ​​corresponding to the at least two prediction results are determined respectively. The prediction result whose plot analysis value is greater than a preset threshold is obtained from the at least two prediction results and used as the prediction derivative video.

7. The method according to claim 6, characterized in that, The process involves inputting the plot logic, the derived guiding information, the reference segments, and the derived plot text into a video prediction network for prediction, generating at least two prediction results, including: The reference segment is input into the style analysis subnetwork in the video prediction network to extract the plot style and determine the plot style information of the reference segment. The plot style information, the derivative guidance information, the plot logic, and the derivative plot text are input into the temporal analysis subnetwork in the video prediction network for temporal analysis to determine at least two predicted events and the time information corresponding to the at least two predicted events respectively. The at least two predicted events and the time information are input into the coherence analysis subnetwork in the video prediction network for event sequence analysis, generating at least two prediction results that satisfy preset sequence conditions.

8. The method according to claim 6, characterized in that, The step of inputting the at least two prediction results into a plot analysis network for plot causal analysis and plot style analysis, and determining the plot analysis values ​​corresponding to the at least two prediction results respectively, includes: The at least two prediction results are input into the causal analysis subnetwork in the plot analysis network to perform plot causal analysis, and the plot causal analysis values ​​corresponding to the at least two prediction results are determined respectively. The at least two prediction results are input into the style matching degree subnetwork in the plot analysis network to perform plot style analysis, and the plot style analysis values ​​corresponding to the at least two prediction results are determined respectively. The plot causal analysis value and the plot style analysis value corresponding to the at least two prediction results are weighted and summed to obtain the plot analysis value corresponding to the at least two prediction results.

9. The method according to claim 1, characterized in that, Before obtaining the derived plot text corresponding to at least one first playback position in the first video, the method further includes: Obtain the video information and script information corresponding to the first video; Based on the video information and the script information, determine the characters, relationships between characters, character descriptions, events, relationships between events, event descriptions, and event-character relationships in the first video; The preset knowledge graph is generated by using the character as a first type of point, the relationship between the characters as an edge between two first type of points, the event as a second type of point, the relationship between the events as an edge between two second type of points, the character description as an attribute representing the first type of point of the character, and the event description as an attribute representing the second type of point of the event, and the event-character relationship as the association between the character represented by the first type of point and the event represented by the second type of point.

10. The method according to claim 9, characterized in that, The character description includes one or more combinations of the following: personality traits, character positioning, and character nickname; The event description includes one or more combinations of the following: the location of the event, the scene of the event, and the time of the event; The relationships between the roles include one or more combinations of the following: adversarial relationship, cooperative relationship; The relationships between events include one or more of the following combinations: causal relationship, parallel relationship.

11. A video playback method, characterized in that, include: When the first video is played to a first playback position, information related to at least one derivative video segment associated with the first playback position is displayed, the derivative video segment being obtained based on the method of any one of claims 1-10; In response to a selection operation of a first relevant information among at least one of the relevant information, a first derived video segment corresponding to the first relevant information is played.

12. The method according to claim 11, characterized in that, Also includes: When the first derived video segment reaches the second playback position, the unplayed content in the first video continues to play, wherein the second playback position is the end position of the first derived video segment.

13. The method according to claim 11, characterized in that, Also includes: When the first derivative video segment is played to the second playback position, the second relevant information of the second derivative video segment in the at least one derivative video segment is displayed, wherein the second playback position is the end position of the first derivative video segment; In response to the selection operation of the second relevant information, the second derived video clip is played.

14. A video processing apparatus, characterized in that, include: The first acquisition module is used to acquire derivative plot text corresponding to at least one first playback position in the first video; The second acquisition module is used to acquire event information and character information associated with the first playback position from the preset knowledge graph corresponding to the first video, wherein the preset knowledge graph is predetermined according to the plot logic of the first video; The first generation module is used to generate derivative guidance information corresponding to the first playback position based on the event information, the character information, and the derivative plot text. The second generation module is used to generate a derivative video segment corresponding to the first playback position based on the plot logic, the derivative guidance information, the reference segment of the first playback position, and the derivative plot text. The derivative video segment does not belong to the original video content of the first video. The reference segment includes the first video content of the first time period before the first playback position and / or the second video content of the second time period after the first playback position.

15. A video playback device, characterized in that, include: The display module is configured to display relevant information of at least one derivative video segment associated with the first playback position when the first video is played to the first playback position, wherein the derivative video segment is obtained based on the method of any one of claims 1-10. The first playback module is configured to play a first derived video segment corresponding to the first relevant information in response to a selection operation of at least one of the relevant information.

16. An electronic device, characterized in that, include: processor; Memory, used to store executable instructions; The processor is configured to retrieve the executable instructions from the memory and execute the executable instructions to implement the method of any one of claims 1-13.

17. A computer-readable storage medium having a computer program stored thereon, characterized in that, The storage medium stores a computer program that, when executed by a processor, causes the processor to implement the method described in any one of claims 1-13.

Images