Resource determination method and apparatus, electronic device, and storage medium

Abstract

The present disclosure provides a resource determination method and device, electronic equipment and storage medium, relates to the technical field of artificial intelligence, in particular to the technical field of deep learning, natural language processing and image generation. The specific implementation scheme is: in response to a resource determination request, at least one target historical video is determined based on a video browsing record of a target object and a behavior log of the target object; a target video currently browsed by the target object is matched with the at least one target historical video based on features, and a matching result is obtained; and based on the matching result, a dynamic graph resource set is determined.

Description

Technical Field

[0001] This disclosure relates to the field of artificial intelligence technology, and in particular to the fields of deep learning, natural language processing, and image generation technology. Background Technology

[0002] With the rapid development of internet technology, various video playback applications offer animated GIF commenting features. Using animated GIFs to post comments allows users to quickly express their opinions, enliven the social atmosphere, and enhance the user experience. Summary of the Invention

[0003] This disclosure provides a resource determination method, apparatus, electronic device, and storage medium.

[0004] According to one aspect of this disclosure, a resource determination method is provided, comprising: in response to a resource determination request, determining at least one target historical video based on the video browsing history of a target object and the behavior log of the target object; performing feature matching between the target video currently being viewed by the target object and the at least one target historical video to obtain a matching result; and determining a set of dynamic graph resources based on the matching result.

[0005] According to another aspect of this disclosure, a resource determination apparatus is provided, comprising: a target historical video determination module, configured to determine at least one target historical video based on the video browsing history of a target object and the behavior log of the target object in response to a resource determination request; a matching result determination module, configured to perform feature matching between the target video currently being viewed by the target object and the at least one target historical video to obtain a matching result; and a resource set determination module, configured to determine a dynamic graph resource set based on the matching result.

[0006] According to another aspect of this disclosure, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the methods described above.

[0007] According to another aspect of this disclosure, a non-transitory computer-readable storage medium is provided storing computer instructions, wherein the computer instructions are used to cause the computer to perform the methods described above.

[0008] According to another aspect of this disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the method described above.

[0009] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0010] The accompanying drawings are provided to better understand this solution and do not constitute a limitation of this disclosure. Wherein:

[0011] Figure 1 This illustration schematically shows an exemplary system architecture to which resource determination methods and apparatus can be applied according to embodiments of the present disclosure;

[0012] Figure 2 A flowchart illustrating a resource determination method according to an embodiment of the present disclosure is shown schematically.

[0013] Figure 3 This illustration schematically shows a flowchart of feature matching between a target video currently being viewed by a target object and a target historical video, according to an embodiment of the present disclosure.

[0014] Figure 4 This illustration schematically shows a flowchart of obtaining target object profile information based on the target object's behavior log according to an embodiment of the present disclosure;

[0015] Figure 5 This schematic diagram illustrates a preset area storing a set of target animated graphs according to an embodiment of the present disclosure;

[0016] Figure 6 A schematic diagram illustrating a resource determination method according to an embodiment of the present disclosure is shown.

[0017] Figure 7 A flowchart illustrating a resource determination method according to another embodiment of this disclosure is shown schematically;

[0018] Figure 8 A block diagram of a resource determination apparatus according to embodiments of the present disclosure is schematically shown; and

[0019] Figure 9 A block diagram of an electronic device suitable for implementing a resource determination method according to an embodiment of the present disclosure is shown schematically. Detailed Implementation

[0020] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0021] In related technologies, video playback applications can provide a dynamic image library, allowing users to select images from the library to post dynamic image comments. The sources of dynamic images in the library include those provided by the video playback application, those shared by other users on the internet, and those collected by users from other users' dynamic image comments. When posting dynamic image comments, users need to search the library for suitable images that express their views or fit the context. However, finding suitable images requires significant manual filtering, and the filtered images may not fully express the user's intent, thus reducing the user's interactive experience.

[0022] In view of this, embodiments of the present disclosure provide a resource determination method that can generate and provide dynamic graph resources that meet the context or intent of the user, thereby enhancing the user's interactive experience. The specific method includes: in response to a resource determination request, determining at least one target historical video based on the target object's video browsing history and the target object's behavior log; performing feature matching between the target video currently viewed by the target object and the at least one target historical video to obtain a matching result; and determining a set of dynamic graph resources based on the matching result.

[0023] Figure 1 The illustration schematically depicts an exemplary system architecture to which resource determination methods and apparatus can be applied according to embodiments of the present disclosure.

[0024] It is important to note that Figure 1 The examples shown are merely examples of system architectures that can be applied to the embodiments of this disclosure, to help those skilled in the art understand the technical content of this disclosure, but do not mean that the embodiments of this disclosure cannot be used in other devices, systems, environments, or scenarios. For example, in another embodiment, an exemplary system architecture to which the resource determination method and apparatus can be applied may include a terminal device, but the terminal device can implement the resource determination method and apparatus provided by the embodiments of this disclosure without interacting with the server.

[0025] like Figure 1 As shown, the system architecture 100 according to this embodiment may include terminal devices 101, 102, and 103, a network 104, and a server 105. The network 104 serves as a medium for providing a communication link between the terminal devices 101, 102, and 103 and the server 105. The network 104 may include various connection types, such as wired and / or wireless communication links, etc.

[0026] Users can use terminal devices 101, 102, and 103 to interact with server 105 via network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, and 103, such as knowledge reading applications, web browser applications, search applications, instant messaging tools, email clients, and / or social platform software, etc. (for example only).

[0027] Terminal devices 101, 102, and 103 can be various electronic devices with displays and web browsing capabilities, including but not limited to smartphones, tablets, laptops, and desktop computers.

[0028] Server 105 can be a server that provides various services.

[0029] It should be noted that the resource determination method provided in this embodiment can generally be executed by terminal devices 101, 102, or 103. Accordingly, the resource determination device provided in this embodiment can also be disposed in terminal devices 101, 102, or 103.

[0030] Alternatively, the resource determination method provided in this embodiment can generally be executed by server 105. Correspondingly, the resource determination device provided in this embodiment can generally be located in server 105. The resource determination method provided in this embodiment can also be executed by a server or server cluster that is different from server 105 and capable of communicating with terminal devices 101, 102, 103 and / or server 105. Correspondingly, the resource determination device provided in this embodiment can also be located in a server or server cluster that is different from server 105 and capable of communicating with terminal devices 101, 102, 103 and / or server 105.

[0031] For example, when a user clicks into the comment section, terminal devices 101, 102, and 103 can trigger a resource determination request, and then send the triggered resource determination request to server 105. Server 105 processes the resource determination request and ultimately determines the set of dynamic image resources. Alternatively, a server or server cluster that can communicate with terminal devices 101, 102, 103, and / or server 105 can process the resource determination request and ultimately determine the set of dynamic image resources.

[0032] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0033] In the technical solution disclosed herein, the collection, storage, use, processing, transmission, provision, disclosure, and application of users' personal information comply with the provisions of relevant laws and regulations, necessary confidentiality measures have been taken, and there is no violation of public order and good morals.

[0034] In the technical solution disclosed herein, the user's authorization or consent is obtained before acquiring or collecting the user's personal information.

[0035] Figure 2 A flowchart illustrating a resource determination method according to an embodiment of this disclosure is shown schematically.

[0036] like Figure 2 As shown, the method includes operations S210 to S230.

[0037] In operation S210, in response to the resource determination request, at least one target historical video is determined based on the target object's video browsing history and the target object's behavior log.

[0038] In operation S220, the target video currently being viewed by the target object is matched with at least one target historical video to obtain the matching result.

[0039] In operation S230, the set of dynamic graph resources is determined based on the matching results.

[0040] According to embodiments of this disclosure, when a target object browses a target video, and after the target object clicks to enter the comment section of the target video, a resource determination request can be triggered.

[0041] According to embodiments of this disclosure, the target object can be a user who browses the target video.

[0042] According to embodiments of this disclosure, when a resource determination request is triggered, the historical videos viewed by the target object can be determined based on the target object's video browsing records within a preset time period, and the operation behavior of the target object for each historical video can be determined based on the target object's behavior log, thereby determining the historical videos that the target object is interested in, i.e., at least one target historical video.

[0043] According to embodiments of this disclosure, the target object's behavior log can record various operations performed by the target object on historical videos, including liking, commenting, sharing, and saving.

[0044] According to embodiments of this disclosure, as an optional implementation, the at least one target historical video can also be determined by incremental filtering. Specifically, at least one first historical video determined when the last resource determination request was triggered can be obtained, and then a second historical video that meets the conditions within the time period between the time when the last resource determination request was triggered and the time when the current resource determination request is triggered can be determined. The set of at least one first historical video and the second historical video is the at least one target historical video.

[0045] According to embodiments of this disclosure, feature matching between the target video currently being viewed by the target object and at least one target historical video can be performed by matching video features extracted from video frame images, audio text, tag information, etc., of the target historical video with video features extracted from video frame images, audio text, tag information, etc., of the target video. Alternatively, feature matching between the target video currently being viewed by the target object and at least one target historical video can also be performed by matching the semantic, contextual, and other information contained in the target historical video with the semantic, contextual, and other information of the target video. Furthermore, feature matching between the target video currently being viewed by the target object and at least one target historical video can also be performed by extracting feature words from both the target video and at least one target historical video, and then performing matching through feature word matching.

[0046] According to embodiments of this disclosure, the obtained matching results may include at least two types of matching results, such as matching and not matching. When the matching result is a match, for example, corpus can be extracted from at least one target historical video, and a motion graph can be generated based on the extracted corpus. The generated motion graphs can be aggregated to obtain a motion graph resource set. When the matching result is a non-match, for example, the motion graph resource set can be determined based on the motion graphs included in the motion graph library, that is, the motion graphs existing in the motion graph library can be added to the motion graph resource set.

[0047] According to embodiments of this disclosure, when a target object browses a target video, a resource determination request can be triggered after the target object clicks to enter the comment section of the target video. When the application's backend responds to this resource determination request, it can determine the videos the target object has previously browsed based on video browsing records, and determine the videos that the target object is interested in from the videos it has previously browsed, i.e., the target historical videos, based on the target object's browsing behavior recorded in the behavior log. By using feature matching, it is possible to obtain a matching result as to whether the target historical videos belong to the target videos currently browsed by the target object. Then, based on the matching result, a set of animated GIF resources to be recommended can be determined. Through the above method, the problem that the animated GIF library cannot meet the personalized needs of the target object can be solved. When the target object needs to comment and interact, animated GIF resources that match the target object's intentions can be provided, reducing the target object's manual filtering operation, thereby effectively improving the target object's interactive experience.

[0048] According to embodiments of this disclosure, determining at least one target historical video based on the target object's video browsing history and the target object's behavior log may include the following operations.

[0049] Based on video browsing history, identify multiple historical videos viewed by the target object within a preset time period; based on behavior logs, identify the log records corresponding to each of the multiple historical videos; and based on the log records corresponding to each of the multiple historical videos, identify at least one target historical video from the multiple historical videos.

[0050] According to the embodiments of this disclosure, the preset time period is not limited, and the preset time period can be set according to the actual situation. For example, the preset time period can be set to the most recent day, the most recent week, the most recent month, or a certain specific time period, etc.

[0051] According to embodiments of this disclosure, based on video browsing history, multiple historical videos viewed by a target object within a preset time period can be identified. Then, log records corresponding to each of the multiple historical videos are determined from the behavior log. These log records may include user actions on the historical videos, such as video navigation, refreshing, liking, commenting, sharing, and saving. If the log record corresponding to a historical video satisfies a target action, then that historical video can be considered the target historical video. The target action can be at least one of the following: liking, commenting, sharing, or saving.

[0052] According to embodiments of this disclosure, by identifying at least one target historical video through multiple historical videos viewed by a target object within a preset time period and the log records corresponding to each of the multiple historical videos, it is possible to accurately determine the historical videos that the user is interested in, thereby meeting the user's personalized needs.

[0053] The following is for reference. Figures 3-6 In conjunction with specific embodiments, Figure 2 The method shown will be further explained.

[0054] Figure 3 The flowchart illustrating feature matching of the target video currently being viewed by the target object with the target historical videos according to an embodiment of the present disclosure is shown schematically.

[0055] like Figure 3 As shown, feature matching between the target video currently being viewed by the target object and the target historical videos includes operations S221 to S222.

[0056] In operation S221, feature extraction is performed on the target video and at least one target historical video respectively to obtain a first set of feature words related to the target video and a second set of feature words related to at least one target historical video.

[0057] In operation S222, feature matching is performed on the first feature word set and the second feature word set to obtain the matching result.

[0058] According to embodiments of this disclosure, when performing feature extraction, feature words can be extracted from video frames, audio text, and tags of the target video and at least one target historical video, respectively, thereby obtaining a first set of feature words related to the target video and a second set of feature words related to at least one target historical video.

[0059] According to embodiments of this disclosure, feature matching of the first feature word set and the second feature word set can be performed by iteratively matching each feature word in the first feature word set with each feature word in the second feature word set. For any feature word in the first feature word set, if that feature word matches a feature word in the second feature word set in terms of textual, semantic, or other elements, then the feature word in the first feature word set is considered to match the feature word in the second feature word set.

[0060] According to embodiments of this disclosure, the matching result can be determined based on the number of matching feature words.

[0061] According to embodiments of this disclosure, the target historical video that matches the target video can be a target historical video that has a contextual association or other relationship with the target video.

[0062] According to embodiments of this disclosure, by performing feature matching on a first set of feature words and a second set of feature words to obtain matching results, the associated target historical videos can be accurately and comprehensively determined for the target video.

[0063] According to embodiments of this disclosure, feature extraction is performed on a target video and at least one target historical video to obtain a first set of feature words related to the target video and a second set of feature words related to at least one target historical video, which may include the following operations.

[0064] For any one of the target video and at least one target historical video, determine at least one feature word related to the video; based on the at least one feature word related to the target video, obtain a first feature word set; and based on at least one feature word related to each of the at least one target historical video, obtain a second feature word set.

[0065] According to embodiments of this disclosure, in the process of determining the target video and at least one target historical video feature word, it is necessary to determine the feature word for each video separately, thereby obtaining at least one feature word related to the target video and at least one feature word related to each of the at least one target historical video.

[0066] According to embodiments of this disclosure, determining at least one feature word associated with a video may include the following operations.

[0067] For a video comprising multiple video frame images, feature images are determined from the multiple video frame images based on the heat analysis results of the multiple video frame images; image recognition is performed on the feature images to determine the first feature word; a second feature word is obtained based on the audio text of the video; a third feature word is obtained based on the tag information of the video; and at least one feature word related to the video is determined based on the first feature word, the second feature word and the third feature word.

[0068] According to embodiments of this disclosure, a video frame image can correspond to a video segment. The video frame image can be determined from multiple video frames included in the video segment. When performing popularity analysis on multiple video frame images included in a video, user behavior data such as likes, comments, favorites, and views corresponding to each video segment of the video can be collected. Based on the collected user behavior data, the popularity of each video segment can be determined through weighted summation or other methods, thereby obtaining the popularity analysis results for each video frame image.

[0069] According to embodiments of this disclosure, a feature image among multiple video frame images can be represented as the video frame image with the highest popularity value indicated by the popularity analysis result. Alternatively, if the popularity analysis result indicates that the video frame image meets a predetermined viewing condition, the video frame image is determined as a feature image, wherein the predetermined viewing condition can be one or more of the following satisfying a predetermined viewing threshold: likes, comments, bullet comments, favorites, view rate, etc.

[0070] According to embodiments of this disclosure, feature matching is performed on a first set of feature words and a second set of feature words to obtain a matching result, which may include the following operations.

[0071] The number of target feature words in the first feature word set that match the feature words in the second feature word set is determined; if the ratio of the number of target feature words to the total number of feature words in the first feature word set is less than a preset value, a matching result indicating that the target video does not match at least one target historical video is obtained; and if the ratio of the number of target feature words to the total number of feature words in the first feature word set is greater than or equal to a preset value, a matching result indicating that the target video matches at least one target historical video is obtained.

[0072] According to embodiments of this disclosure, the matching result can be determined by the number of target feature words in the first feature word set that match the feature words in the second feature word set.

[0073] According to the embodiments of this disclosure, the preset value is not limited, and a suitable preset value can be determined according to the actual situation. For example, the preset value can be set to 0.8, 0.9, etc.

[0074] According to embodiments of this disclosure, determining a set of dynamic graph resources based on matching results may include the following operations.

[0075] If the matching result indicates that the target video matches at least one target historical video, multiple hotspot segments are identified from the at least one target historical video based on the matching result; multiple target animated graphs are generated based on the multiple hotspot segments; and an animated graph resource set is obtained based on the multiple target animated graphs.

[0076] According to embodiments of this disclosure, trending segments can be determined from target historical videos. One or more trending segments can be determined from multiple video segments included in each target historical video. A video segment can be a series of consecutive video frames. Based on user behavior data such as likes, comments, favorites, and views for each video segment in the target historical video, the popularity of each video segment is determined, and one or more video segments with the highest popularity are selected as trending segments.

[0077] According to embodiments of this disclosure, one or more target animated images can be generated based on each hotspot segment. The processing operations on the hotspot segments may include, but are not limited to, resolution adjustment, size adjustment, duration trimming, and adding special effects.

[0078] According to embodiments of this disclosure, a set of dynamic graph resources can be obtained by aggregating and collecting multiple target dynamic graphs.

[0079] According to embodiments of this disclosure, when the matching result indicates that the target video matches at least one target historical video, multiple hotspot segments can be determined from the at least one target historical video that matches the target video, and the hotspot frequency bands can be processed to obtain a target dynamic map, thereby generating a target dynamic map that conforms to the user's intent.

[0080] According to embodiments of this disclosure, generating multiple target dynamic graphs based on multiple hotspot segments may include the following operations.

[0081] For hotspot segments, the hotspot segments are cropped based on preset duration and preset size to generate at least one target animation related to the hotspot segments; and multiple target animations are obtained based on at least one target animation related to each of the multiple hotspot segments.

[0082] According to embodiments of this disclosure, hotspot segments can be cropped according to the standard of dynamic graphs, i.e., based on preset time and preset size, thereby generating a more standardized target dynamic graph.

[0083] According to embodiments of this disclosure, furthermore, after the cropping operation, operations such as format conversion can be performed to convert the cropped video clips into a target animated image.

[0084] According to embodiments of this disclosure, generating multiple target dynamic graphs based on multiple hotspot segments may include the following operations.

[0085] For hotspot segments, identify the character image information related to the hotspot segments; based on the character image information, identify at least one target animated image related to the character image information from multiple animated images included in the animated image resource library; and based on at least one target animated image related to each of the multiple hotspot segments, obtain multiple target animated images.

[0086] According to embodiments of this disclosure, as an optional implementation, a target dynamic image can also be obtained based on the character image information, which contains the task information. For example, face recognition can be performed on each video frame in a hotspot segment to determine the character image information of the character included in the hotspot segment. Using this character image information, at least one target dynamic image related to the character image information can be determined from multiple dynamic images included in the dynamic image resource library.

[0087] According to the embodiments of this disclosure, the character image information is not limited. The character image information can be any information that can represent the character image, such as: facial features, character name, performer name, etc.

[0088] According to embodiments of this disclosure, determining a set of dynamic graph resources based on matching results may include the following operations.

[0089] If the matching result indicates that the target video does not match at least one target historical video, a set of dynamic graph resources is determined from the dynamic graph resource library based on the profile information of the target object.

[0090] According to embodiments of this disclosure, when the matching result indicates that the target video does not match at least one target historical video, a set of dynamic graph resources corresponding to the portrait information of the target object can be determined based on the portrait information of the target object.

[0091] According to embodiments of this disclosure, the profile information can be determined based on the target object's historical behavior. Specifically, the target object profile can be determined based on the comment information containing animated GIFs previously sent by the target object, or the profile information of the target object can be determined based on the types of videos watched by the target object during a preset time period and the behavior logs for different types of videos. Similarly, the user profile can also be determined by other methods.

[0092] According to embodiments of this disclosure, when the matching result indicates that the target video does not match at least one target historical video, a set of dynamic graph resources that matches the intent of the target object can be determined from the dynamic graph resource set in the dynamic graph resource library based on the profile information of the target object.

[0093] According to embodiments of this disclosure, determining a set of dynamic graph resources from a dynamic graph resource library based on the profile information of the target object may include the following operations.

[0094] Based on the profile information of the target object, the priority weights of the multiple dynamic graphs included in the dynamic graph resource library are determined; based on the priority weights of the multiple dynamic graphs, multiple target dynamic graphs are determined from the multiple dynamic graphs; and based on the multiple target dynamic graphs, a set of dynamic graph resources is obtained.

[0095] According to embodiments of this disclosure, the profile information of the target object includes the target object's emotional tendencies, the target object's interests and preferences, the event information that the target object is interested in, and so on.

[0096] According to embodiments of this disclosure, based on the profile information of the target object, the priority weight of each of the multiple dynamic images included in the dynamic image resource library can be determined. Specifically, the interests and preferences of the target object can be determined based on the profile information. For example, if the target object likes food, the priority weight of dynamic images containing food can be increased.

[0097] According to embodiments of this disclosure, the priority weight of each animated image can be determined by identifying recent events of interest to the target object based on the profile information. For example, if the target object is currently watching a live stream of an event and the target video being watched is related to that event, the priority weight of the animated image related to that event can be increased.

[0098] According to embodiments of this disclosure, the emotional tendency of the target object can also be determined based on the target object's profile information. For example, if a user's recent emotional tendency is positive, the priority weight of the dynamic image representing positive emotion can be increased.

[0099] According to embodiments of this disclosure, each dynamic graph is sorted based on its respective priority weight, thereby selecting the dynamic graph that satisfies the predetermined sorting position as the target dynamic graph.

[0100] According to embodiments of this disclosure, factors such as the geographical location, time, and weather of the target object can also be considered. If the target video currently being viewed by the target object is related to the above factors, a dynamic image corresponding to the above factors can be determined from the dynamic image resource library, and the weight and priority of the dynamic image can be increased.

[0101] According to embodiments of this disclosure, the dynamic graphs commonly used by the target object's friends, followers, and other objects can also be determined by analyzing the target object's social network relationships, and the weight and priority of the dynamic graphs can be increased.

[0102] According to embodiments of this disclosure, the priority weights of multiple animated images included in the animated image resource library are determined by using the profile information of the target object. This reduces the number of situations where there are no animated images that match the target object's intent, and eliminates the need for the target object to search through a large number of animated images, thus achieving a better technical effect of meeting user needs.

[0103] According to embodiments of this disclosure, a profile of the target object is obtained based on the target object's behavior logs.

[0104] According to embodiments of this disclosure, the profile information of the target object can also be obtained through the target object's behavior log, that is, the profile information of the target object can be determined through the target object's historical comment information, historical like information, etc.

[0105] According to embodiments of this disclosure, different target objects may correspond to different profile information. Based on the behavior logs of the target object, the profile information corresponding to the target object can be determined, and a dynamic graph resource set that better matches the intention of the target object can be generated based on the profile information.

[0106] Figure 4The flowchart illustrating the process of obtaining target object profile information based on the target object's behavior log according to an embodiment of the present disclosure is shown.

[0107] like Figure 4 As shown, based on the behavior log of the target object, the profile information of the target object can be obtained, which may include operations S410 to S450.

[0108] When operating S410, multiple historical comment messages of the target object are obtained based on the target object's behavior log.

[0109] When operating S420, multiple historical comment messages are input into the dynamic image recognition model to obtain the recognition results for each historical comment message.

[0110] In operation S430, based on the identification results of multiple historical comment information, at least one target comment information containing a dynamic image is determined from the multiple historical comment information.

[0111] In operation S440, the text information contained in at least one target comment is input into the sentiment analysis model to obtain the sentiment analysis results of the target object.

[0112] When operating the S450, based on the sentiment analysis results, the profile information of the target object is obtained.

[0113] According to embodiments of this disclosure, inputting historical comment information into a dynamic image recognition model can determine whether each historical comment information includes a dynamic image. Inputting the file information included in the target comment information that includes a dynamic image into a sentiment analysis model can obtain the sentiment analysis results of the target object, thereby obtaining the profile information of the target object based on the sentiment analysis results.

[0114] According to embodiments of this disclosure, the motion graph recognition model and the sentiment analysis model are trained using labeled historical comment information. The labels on the historical comment information may include information indicating whether the historical comment information contains a motion graph, and information indicating the sentiment expressed by the historical comment information.

[0115] According to embodiments of this disclosure, during the training process of labeled historical comment information, the animated images in the historical comment information can be decomposed into frame images, and the features of each frame image can be extracted. The method of extracting the features of the frame images is not limited and can be any deep learning or other method that can extract the features of the frame images. For example, it can be to use a convolutional neural network to extract image features and represent each frame as a vector.

[0116] According to embodiments of this disclosure, the feature representations of the extracted frame images are arranged in chronological order to form a time-series data sequence.

[0117] According to embodiments of this disclosure, a self-attention mechanism is employed to encode and model temporal data sequences. The feature representation of each frame undergoes multiple attention calculations and feature fusions, thereby enabling better learning of temporal dependencies and relationships. Therefore, the self-attention mechanism can better capture the relationships and importance between different frames in a temporal data sequence. Specifically, a temporal encoder can be constructed using a self-attention mechanism, and the specific method of the self-attention mechanism employed is not limited; for example, the attention mechanism in Transformer can be used.

[0118] According to embodiments of this disclosure, the encoded temporal data sequence can be used to represent the entire dynamic graph, and the representation takes into account the temporal relationship between different frame images.

[0119] According to embodiments of this disclosure, dynamic graphs are identified by encoding and representing time-series data sequences, followed by the use of fully connected layers or other classification models. By learning key information from the time-series data sequences, the model can more accurately identify dynamic graphs.

[0120] According to embodiments of this disclosure, when implementing a self-attention mechanism, multi-head attention and positional encoding can be used to enhance the dynamic graph recognition model's attention to different frame images in time-series data.

[0121] According to embodiments of this disclosure, dynamic graph recognition using a self-attention mechanism avoids the traditional method of processing images frame by frame. By performing self-attention encoding on the temporal data sequence, the dynamic graph recognition model can better capture the relationships between frame images, thereby achieving more accurate and efficient recognition of the entire dynamic graph and improving the efficiency and performance of dynamic graph recognition.

[0122] According to embodiments of this disclosure, the sentiment analysis model may include a sentiment confusion detection model, a sentiment classification model, and a traditional sentiment analysis model.

[0123] According to embodiments of this disclosure, the sentiment confusion detection model can be a binary classification model used to determine whether text information contains sentiment confusion. The specific model used for the binary classification model is not limited; it can be any model capable of performing sentiment confusion detection, such as a convolutional neural network or a recurrent neural network. When training the sentiment confusion detection model, samples can be divided into negative samples or positive samples, where negative samples are labeled sentiment-confusing text, and positive samples are unconfusing text.

[0124] According to embodiments of this disclosure, text information detected as having conflicting emotions can be input into a sentiment classification model to identify the specific emotions present in the confusion. The specific sentiment classification model used is not limited; an appropriate model can be selected based on the implementation requirements. Furthermore, the number of sentiment categories in the sentiment classification model is not limited; it can be set according to the actual number of sentiment categories.

[0125] According to embodiments of this disclosure, text information detected as having no emotional confusion can be input into a traditional sentiment analysis model for sentiment classification. The specific traditional sentiment analysis model used is not limited; an appropriate model can be selected based on the implementation requirements. Furthermore, the number of sentiment categories in the traditional sentiment analysis model is not limited; it can be set according to the actual number of sentiment categories.

[0126] According to embodiments of this disclosure, the sentiment analysis model may further include a sentiment graph analysis model.

[0127] According to embodiments of this disclosure, a sentiment relationship graph can be established by analyzing the textual information contained in each target comment based on a sentiment graph analysis model. Nodes represent different sentiments, and edges represent the relationships between sentiments. The relationships between sentiments can include similarity, transformation relationships, etc.

[0128] According to embodiments of this disclosure, the specific method used in the sentiment analysis model is not limited, and it can be a graph neural network (GNN).

[0129] According to embodiments of this disclosure, based on the sentiment analysis results, profile information of the target object can be obtained, and the sentiment analysis results may include the target object's emotional tendencies, etc.

[0130] According to embodiments of this disclosure, the text information currently entered by the target object in the comment section can also be monitored in real time, and the real-time sentiment tendency of the target object can be determined based on this text information. This real-time sentiment tendency is then weighted and processed with previously determined historical sentiment tendencies to more accurately determine the target object's sentiment tendency.

[0131] According to embodiments of this disclosure, a decay factor can be set for multiple historical comment information of a target object, so that the influence of earlier historical data gradually decreases during the recommendation process, thereby focusing more on recent sentiment tendencies.

[0132] According to embodiments of this disclosure, reinforcement learning methods can also be introduced to dynamically adjust the recommendation strategy of the aforementioned dynamic graph to maximize the satisfaction of user sentiment. For example, algorithms such as Q-learning can be used to continuously adjust the recommendation weight and priority of the dynamic graph based on the user's real-time sentiment and historical data.

[0133] According to embodiments of this disclosure, multiple historical comment information of a target object is input into a dynamic graph recognition model, and the text information contained in at least one target comment information is input into a sentiment analysis model to obtain the sentiment analysis result of the target object. Based on the sentiment analysis result, the profile information of the target object is obtained. This allows for more accurate and faster analysis of the historical comment information of the target object, thereby obtaining a more accurate profile information that represents the target object.

[0134] According to embodiments of this disclosure, the resource determination method may further include the following operations.

[0135] In response to a comment input request, multiple target animated GIFs from the animated GIF resource collection are displayed in a preset area.

[0136] Figure 5 A schematic diagram illustrating a preset area storing a set of target dynamic graphs according to an embodiment of the present disclosure is shown.

[0137] like Figure 5 As shown, when the target user's client device is a mobile phone, and the target user intends to post a comment, they can select a target animated image that meets their requirements from the preset area. The target animated image will appear after "User Input:". The preset area includes multiple target animated images.

[0138] According to the embodiments of this disclosure, the layout and design of the preset area and the comment posting area, the number of target animated images, and the arrangement method are only illustrative. Different layouts and designs of the preset area and the comment posting area, the number of target animated images, and the arrangement method can be adopted according to actual needs.

[0139] According to embodiments of this disclosure, after determining the set of animated graph resources corresponding to the target object, in response to a comment input request, multiple target animated graphs included in the animated graph set can be displayed in a preset area so that the target object can post a comment based on the multiple target animated graphs included in the animated graph resource set.

[0140] According to embodiments of this disclosure, a comment input request can be for a user to open the comment area or for a user to click on a preset area, etc.

[0141] According to embodiments of this disclosure, multiple target animations may also have their own indexes, and the target animation corresponding to the index can be found in a preset area based on the index of the target animation.

[0142] Figure 6 A schematic diagram of a resource determination method according to an embodiment of the present disclosure is shown.

[0143] According to embodiments of this disclosure, such as Figure 6As shown, in response to the resource determination request, multiple historical videos 603 viewed by the target object within a preset time period are determined through video browsing records 601. Based on behavior logs 602, log records 604 corresponding to each of the multiple historical videos are determined. Using log records 604, one or more target historical videos 605 that meet the target operation behavior are determined from the multiple historical videos 603.

[0144] According to embodiments of this disclosure, feature extraction is performed on the target video 607 currently being viewed by the target object to obtain a first feature word set 606 related to the target video. Feature extraction is performed on the target historical video 605 to obtain a second feature word set 608 related to at least one target historical video 605. Feature matching is performed between the first feature word set 606 and the second feature word set 608 to obtain a matching result 609.

[0145] According to embodiments of this disclosure, when a matching result 609 indicates that a target video 606 matches at least one target historical video 605, a plurality of hotspot segments 610 are determined from the at least one target historical video 605 using the matching result 609. A plurality of first target animated graphs 611 are generated using the plurality of hotspot segments 610, thereby obtaining a first animated graph resource set 612.

[0146] According to an embodiment of this disclosure, when the matching result 609 indicates that the target video 606 does not match at least one target historical video, the priority weight 614 of each of the multiple animated images included in the animated image resource library is determined using the portrait information 613 of the target object. Based on the priority weight 614, multiple second target animated images 615 are determined from the multiple animated images, thereby obtaining a second animated image resource set 616 based on the multiple second target animated images 615.

[0147] Figure 7 A flowchart illustrating a resource determination method according to another embodiment of this disclosure is shown schematically.

[0148] like Figure 7 As shown, the method includes operations S701 to S710.

[0149] In operation S701, if it is detected that the target object clicks into the comment area of ​​the target video currently being viewed by the target object, a resource determination request is triggered.

[0150] In operation S702, in response to a resource determination request, at least one target historical video is determined based on the target object's video browsing history and the target object's behavior log.

[0151] In operation S703, a first set of feature words for the target video and a second set of feature words for each of at least one target historical video are determined.

[0152] In operation S704, it is detected whether the first feature word set and the second feature word set match, and the matching results of the target video and at least one target historical video are obtained respectively.

[0153] In operation S705, if the matching result indicates that the target video matches at least one target historical video, multiple hotspot segments are identified from at least one target historical video based on the matching result.

[0154] When operating the S706, multiple target dynamic graphs are generated based on multiple hotspot segments.

[0155] When operating the S707, a set of dynamic graph resources is obtained based on multiple target dynamic graphs.

[0156] When operating S708, multiple target animated graphs included in the animated graph resource set are displayed in a preset area so that the target object can comment based on the multiple target animated graphs.

[0157] In operation S709, if the matching result indicates that the target video does not match at least one target historical video, the priority weight of each of the multiple animated graphs included in the animated graph resource library is determined based on the portrait information of the target object.

[0158] In operation S710, multiple target dynamic graphs are determined from the multiple dynamic graphs based on their respective priority weights, and operation S706 is returned.

[0159] Figure 8 A block diagram of a resource determination apparatus according to an embodiment of the present disclosure is shown schematically.

[0160] like Figure 8 As shown, the resource determination device 800 includes: a target historical video determination module 810, a matching result determination module 820, and a resource set determination module 830.

[0161] The target historical video determination module 810 is used to determine at least one target historical video in response to a resource determination request, based on the target object's video browsing history and the target object's behavior log.

[0162] The matching result determination module 820 is used to perform feature matching between the target video currently being viewed by the target object and at least one target historical video to obtain the matching result.

[0163] The resource set determination module 830 is used to determine the dynamic graph resource set based on the matching results.

[0164] According to embodiments of this disclosure, the matching result determination module 820 includes: a feature word set determination submodule and a feature matching submodule.

[0165] The feature word set determination submodule is used to extract features from the target video and at least one target historical video respectively, to obtain a first feature word set related to the target video and a second feature word set related to at least one target historical video.

[0166] The feature matching submodule is used to perform feature matching on the first feature word set and the second feature word set to obtain the matching result.

[0167] According to embodiments of this disclosure, the feature word set determination submodule includes: a feature word determination unit, a first feature word set determination unit, and a second feature word set determination unit.

[0168] The feature word determination unit is used to determine at least one feature word related to the target video and at least one target historical video for any one of the target video and the target historical video.

[0169] The first feature word set determination unit is used to obtain a first feature word set based on at least one feature word related to the target video.

[0170] The second feature word set determination unit is used to obtain a second feature word set based on at least one feature word associated with each of at least one target historical video.

[0171] According to embodiments of this disclosure, the feature word determination unit includes: a feature image determination subunit, a first feature word determination subunit, a second feature word determination subunit, a third feature word determination subunit, and a feature word determination subunit.

[0172] The feature image determination subunit is used to determine feature images from multiple video frame images based on the heat analysis results of the multiple video frame images included in the video.

[0173] The first feature word determination subunit is used to perform image recognition on the feature image in order to determine the first feature word.

[0174] The second feature word determination subunit is used for video-based audio text to obtain the second feature word.

[0175] The third feature word determination subunit is used to obtain the third feature word based on the video's tag information.

[0176] The feature word determination subunit is used to determine at least one feature word related to the video based on the first feature word, the second feature word, and the third feature word.

[0177] According to embodiments of this disclosure, the feature matching submodule includes: a quantity determination unit, a first matching result determination unit, and a second matching result determination unit.

[0178] The quantity determination unit is used to determine the number of target feature words in the first feature word set that match the feature words in the second feature word set.

[0179] The first matching result determination unit is used to obtain a matching result indicating that the target video does not match at least one target historical video when the ratio of the number of target feature words to the total number of feature words included in the first feature word set is less than a preset value.

[0180] The second matching result determination unit is used to obtain a matching result indicating that the target video matches at least one target historical video when the ratio of the number of target feature words to the total number of feature words included in the first feature word set is greater than or equal to a preset value.

[0181] According to embodiments of this disclosure, the resource set determination module 830 includes: a hotspot segment determination submodule, a target dynamic graph generation submodule, and a first resource set determination submodule.

[0182] The hotspot segment identification submodule is used to identify multiple hotspot segments from at least one target historical video based on the matching results, provided that the matching results indicate that the target video matches at least one target historical video.

[0183] The target animation generation submodule is used to generate multiple target animations based on multiple hotspot segments.

[0184] The first resource set determination submodule is used to obtain a dynamic graph resource set based on multiple target dynamic graphs.

[0185] According to embodiments of this disclosure, the target dynamic graph generation submodule includes: a hotspot segment processing unit and a first target dynamic graph determination unit.

[0186] The hot spot segment processing unit is used to crop the hot spot segments based on a preset duration and a preset size, and generate at least one target animation related to the hot spot segments.

[0187] The first target dynamic graph determination unit is used to obtain multiple target dynamic graphs based on at least one target dynamic graph associated with each of the multiple hotspot segments.

[0188] According to embodiments of this disclosure, the target dynamic graph generation submodule includes: an information determination unit, a dynamic graph determination unit, and a second target dynamic graph determination unit.

[0189] The information determination unit is used to determine the image information of people related to the hot spot segments.

[0190] The dynamic image determination unit is used to determine at least one target dynamic image related to the character image information from multiple dynamic images included in the dynamic image resource library, based on the character image information.

[0191] The second target dynamic graph determination unit is used to obtain multiple target dynamic graphs based on at least one target dynamic graph associated with each of the multiple hotspot segments.

[0192] According to an embodiment of this disclosure, the resource set determination module 830 includes: a second resource set determination submodule.

[0193] The second resource set determination submodule is used to determine the dynamic graph resource set from the dynamic graph resource library based on the portrait information of the target object when the matching result indicates that the target video does not match at least one target historical video.

[0194] According to embodiments of this disclosure, the second resource set determination submodule includes: a priority weight determination unit, a third target dynamic graph determination unit, and a resource set determination unit.

[0195] The priority weight determination unit is used to determine the priority weight of each of the multiple dynamic graphs included in the dynamic graph resource library based on the portrait information of the target object.

[0196] The third target dynamic graph determination unit is used to determine multiple target dynamic graphs from multiple dynamic graphs based on their respective priority weights.

[0197] The resource set determination unit is used to obtain a dynamic graph resource set based on multiple target dynamic graphs.

[0198] According to embodiments of this disclosure, the resource determination device further includes: a portrait information determination module.

[0199] The profile information determination module is used to obtain profile information of the target object based on the target object's behavior logs.

[0200] According to embodiments of this disclosure, the profile information determination module includes: a historical comment information determination submodule, a recognition result determination submodule, a target comment information determination submodule, a sentiment analysis result determination submodule, and a profile information determination submodule.

[0201] The historical comment information determination submodule is used to obtain multiple historical comment information of the target object based on the target object's behavior log.

[0202] The identification result determination submodule is used to input multiple historical comment information into the dynamic graph recognition model to obtain the identification results of each historical comment information.

[0203] The target comment information determination submodule is used to determine at least one target comment containing an animated image from multiple historical comment information based on the recognition results of each of the multiple historical comment information.

[0204] The sentiment analysis result determination submodule is used to input the text information contained in at least one target comment into the sentiment analysis model to obtain the sentiment analysis result of the target object.

[0205] The profile information determination submodule is used to obtain profile information of the target object based on the sentiment analysis results.

[0206] According to embodiments of this disclosure, the target historical video determination module 810 includes: a historical video determination submodule, a log recording determination submodule, and a target historical video determination submodule.

[0207] The historical video identification submodule is used to identify multiple historical videos viewed by a target object within a preset time period based on video browsing history.

[0208] The log recording determination submodule is used to determine the log records corresponding to each of the multiple historical videos based on the behavior logs.

[0209] The target historical video determination submodule is used to determine at least one target historical video from multiple historical videos based on the log records corresponding to each of the multiple historical videos.

[0210] According to embodiments of this disclosure, the resource determination apparatus further includes a display module.

[0211] The display module is used to respond to comment input requests and display multiple target animated images from the animated image resource collection in a preset area.

[0212] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0213] According to an embodiment of the present disclosure, an electronic device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the methods described above.

[0214] According to embodiments of this disclosure, a non-transitory computer-readable storage medium stores computer instructions, wherein the computer instructions are used to cause a computer to perform the methods described above.

[0215] According to an embodiment of this disclosure, a computer program product includes a computer program that, when executed by a processor, implements the method described above.

[0216] Figure 9 A schematic block diagram of an example electronic device 900 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0217] like Figure 9 As shown, device 900 includes a computing unit 901, which can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) 902 or a computer program loaded from storage unit 908 into random access memory (RAM) 903. RAM 903 may also store various programs and data required for the operation of device 900. The computing unit 901, ROM 902, and RAM 903 are interconnected via bus 904. Input / output (I / O) interface 905 is also connected to bus 904.

[0218] Multiple components in device 900 are connected to input / output (I / O) interface 905, including: input unit 906, such as keyboard, mouse, etc.; output unit 907, such as various types of monitors, speakers, etc.; storage unit 908, such as disk, optical disk, etc.; and communication unit 909, such as network card, modem, wireless transceiver, etc. Communication unit 909 allows device 900 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0219] The computing unit 901 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 901 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 901 performs the various methods and processes described above, such as the resource determination method. For example, in some embodiments, the resource determination method may be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 908. In some embodiments, part or all of the computer program may be loaded and / or installed on device 900 via ROM 902 and / or communication unit 909. When the computer program is loaded into RAM 903 and executed by the computing unit 901, one or more steps of the resource determination method described above may be performed. Alternatively, in other embodiments, the computing unit 901 may be configured to perform the resource determination method by any other suitable means (e.g., by means of firmware).

[0220] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0221] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0222] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0223] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0224] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with embodiments of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0225] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, distributed system servers, or servers incorporating blockchain technology.

[0226] It should be understood that the various forms of processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0227] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A resource determination method, comprising: In response to a resource determination request, at least one target historical video is determined based on the target object's video browsing history and the target object's behavior log; The target video currently viewed by the target object is matched with the at least one target historical video to obtain a matching result; as well as Based on the matching results, a set of dynamic graph resources is determined.

2. The method according to claim 1, wherein, The step of performing feature matching between the target video currently viewed by the target object and the at least one target historical video to obtain a matching result includes: Feature extraction is performed on the target video and the at least one target historical video respectively to obtain a first set of feature words related to the target video and a second set of feature words related to the at least one target historical video; and The first set of feature words and the second set of feature words are matched to obtain the matching result.

3. The method according to claim 2, wherein, The step of extracting features from the target video and the at least one target historical video to obtain a first set of feature words related to the target video and a second set of feature words related to the at least one target historical video includes: For any one of the target video and the at least one target historical video, determine at least one feature word related to the video; The first feature word set is obtained based on at least one feature word associated with the target video; and The second set of feature words is obtained based on at least one feature word associated with each of the at least one target historical video.

4. The method according to claim 3, wherein, Determining at least one feature word related to the video includes: For the video comprising multiple video frame images, feature images are determined from the multiple video frame images based on the heat analysis results of the multiple video frame images; Image recognition is performed on the feature image to determine the first feature word; Based on the audio text of the video, the second feature word is obtained; Based on the tag information of the video, a third feature word is obtained; and Based on the first feature word, the second feature word, and the third feature word, at least one feature word related to the video is determined.

5. The method according to claim 2, wherein, The step of performing feature matching between the first feature word set and the second feature word set to obtain the matching result includes: Determine the number of target feature words in the first feature word set that match the feature words in the second feature word set; If the ratio of the number of target feature words to the total number of feature words included in the first feature word set is less than a preset value, a matching result indicating that the target video does not match the at least one target historical video is obtained; and If the ratio of the number of target feature words to the total number of feature words included in the first feature word set is greater than or equal to the preset value, the matching result indicating that the target video matches the at least one target historical video is obtained.

6. The method according to claim 1, wherein, The process of determining the dynamic graph resource set based on the matching results includes: If the matching result indicates that the target video matches the at least one target historical video, then based on the matching result, multiple hotspot segments are determined from the at least one target historical video. Based on the aforementioned multiple hotspot segments, multiple target dynamic graphs are generated; and Based on the multiple target dynamic graphs, the dynamic graph resource set is obtained.

7. The method according to claim 6, wherein, The generation of multiple target dynamic graphs based on the multiple hotspot segments includes: For the hotspot segment, the hotspot segment is cropped based on a preset duration and a preset size to generate at least one target animated image related to the hotspot segment; and The plurality of target dynamic graphs are obtained based on at least one target dynamic graph associated with each of the plurality of hotspot segments.

8. The method according to claim 6, wherein, The generation of multiple target dynamic graphs based on the multiple hotspot segments includes: For the aforementioned trending segments, determine the character information related to the trending segments; Based on the character image information, at least one target animated image related to the character image information is determined from multiple animated images included in the animated image resource library; and The plurality of target dynamic graphs are obtained based on at least one target dynamic graph associated with each of the plurality of hotspot segments.

9. The method according to claim 1, wherein, The process of determining the dynamic graph resource set based on the matching results includes: If the matching result indicates that the target video does not match the at least one target historical video, the set of dynamic graph resources is determined from the dynamic graph resource library based on the portrait information of the target object.

10. The method according to claim 9, wherein, The step of determining the dynamic graph resource set from the dynamic graph resource library based on the portrait information of the target object includes: Based on the profile information of the target object, the priority weights of the multiple dynamic images included in the dynamic image resource library are determined. Based on the priority weights of the multiple dynamic graphs, multiple target dynamic graphs are determined from the multiple dynamic graphs; and Based on the multiple target dynamic graphs, the dynamic graph resource set is obtained.

11. The method according to claim 9 or 10, further comprising: Based on the target object's behavior logs, the target object's profile information is obtained.

12. The method according to claim 11, wherein, The process of obtaining the profile information of the target object based on its behavior logs includes: Based on the target object's behavior logs, multiple historical comment records of the target object are obtained; The multiple historical comment records are input into the dynamic image recognition model to obtain the recognition results for each of the multiple historical comment records; Based on the identification results of each of the multiple historical comment information, at least one target comment information containing a dynamic image is determined from the multiple historical comment information; The text information contained in each of the at least one target comment is input into the sentiment analysis model to obtain the sentiment analysis result of the target object; and Based on the sentiment analysis results, the profile information of the target object is obtained.

13. The method according to claim 1, wherein, The determination of at least one target historical video based on the target object's video browsing history and the target object's behavior log includes: Based on the video browsing history, determine multiple historical videos viewed by the target object within a preset time period; Based on the behavior logs, determine the log records corresponding to each of the multiple historical videos; and Based on the log records corresponding to each of the plurality of historical videos, at least one target historical video is determined from the plurality of historical videos.

14. The method according to claim 1, further comprising: In response to a comment input request, multiple target animated images included in the animated image resource set are displayed in a preset area.

15. A resource determination apparatus, comprising: The target historical video determination module is used to determine at least one target historical video in response to a resource determination request, based on the target object's video browsing history and the target object's behavior log. The matching result determination module is used to perform feature matching between the target video currently viewed by the target object and the at least one target historical video to obtain a matching result; as well as The resource set determination module is used to determine the dynamic graph resource set based on the matching results.

16. The apparatus according to claim 15, wherein, The matching result determination module includes: A feature word set determination submodule is used to extract features from the target video and the at least one target historical video, respectively, to obtain a first feature word set related to the target video and a second feature word set related to the at least one target historical video; and The feature matching submodule is used to perform feature matching between the first feature word set and the second feature word set to obtain the matching result.

17. The apparatus according to claim 16, wherein, The feature word set determination submodule includes: The feature word determination unit is used to determine at least one feature word related to the target video and the at least one target historical video for any one of the target video and the at least one target historical video; The first feature word set determination unit is configured to obtain the first feature word set based on at least one feature word related to the target video; and The second feature word set determination unit is used to obtain the second feature word set based on at least one feature word associated with each of the at least one target historical video.

18. The apparatus according to claim 17, wherein, The feature word determination unit includes: The feature image determination subunit is used to determine feature images from the multiple video frame images included in the video based on the heat analysis results of the multiple video frame images; The first feature word determination subunit is used to perform image recognition on the feature image to determine the first feature word; The second feature word determination subunit is used to obtain the second feature word based on the audio text of the video; The third feature word determination subunit is used to obtain the third feature word based on the tag information of the video; and The feature word determination subunit is used to determine at least one feature word related to the video based on the first feature word, the second feature word, and the third feature word.

19. The apparatus according to claim 16, wherein, The feature matching submodule includes: A quantity determination unit is used to determine the number of target feature words in the first feature word set that match the feature words in the second feature word set. The first matching result determining unit is configured to, when the ratio of the number of target feature words to the total number of feature words included in the first feature word set is less than a preset value, obtain a matching result indicating that the target video does not match the at least one target historical video; and The second matching result determination unit is used to obtain the matching result indicating that the target video matches the at least one target historical video when the ratio of the number of target feature words to the total number of feature words included in the first feature word set is greater than or equal to the preset value.

20. The apparatus according to claim 15, wherein, The resource set determination module includes: The hotspot segment determination submodule is used to determine multiple hotspot segments from the at least one target historical video based on the matching result when the matching result indicates that the target video matches the at least one target historical video. The target dynamic graph generation submodule is used to generate multiple target dynamic graphs based on the multiple hotspot segments; and The first resource set determination submodule is used to obtain the dynamic graph resource set based on the multiple target dynamic graphs.

21. The apparatus according to claim 20, wherein, The target dynamic graph generation submodule includes: A hotspot segment processing unit is configured to crop the hotspot segment based on a preset duration and a preset size, and generate at least one target animated image related to the hotspot segment; and The first target dynamic graph determination unit is used to obtain the plurality of target dynamic graphs based on at least one target dynamic graph associated with each of the plurality of hotspot segments.

22. The apparatus according to claim 20, wherein, The target dynamic graph generation submodule includes: The information determination unit is used to determine, for the hotspot segment, the image information of the person related to the hotspot segment; An animated image determination unit is configured to, based on the character image information, determine at least one target animated image related to the character image information from multiple animated images included in the animated image resource library; and The second target dynamic graph determination unit is used to obtain the plurality of target dynamic graphs based on at least one target dynamic graph associated with each of the plurality of hotspot segments.

23. The apparatus according to claim 15, wherein, The resource set determination module includes: The second resource set determination submodule is used to determine the dynamic graph resource set from the dynamic graph resource library based on the portrait information of the target object when the matching result indicates that the target video does not match the at least one target historical video.

24. The apparatus according to claim 23, wherein, The second resource set determination submodule includes: The priority weight determination unit is used to determine the priority weight of each of the multiple dynamic images included in the dynamic image resource library based on the portrait information of the target object. The third target dynamic graph determination unit is used to determine multiple target dynamic graphs from the multiple dynamic graphs based on their respective priority weights; and The resource set determination unit is used to obtain the dynamic graph resource set based on the multiple target dynamic graphs.

25. The apparatus according to claim 23 or 24, further comprising: The profile information determination module is used to obtain the profile information of the target object based on the target object's behavior log.

26. The apparatus according to claim 25, wherein, The portrait information determination module includes: The historical comment information determination submodule is used to obtain multiple historical comment information of the target object based on the target object's behavior log; The identification result determination submodule is used to input the multiple historical comment information into the dynamic graph recognition model to obtain the identification results of each of the multiple historical comment information; The target comment information determination submodule is used to determine at least one target comment information containing a dynamic image from the multiple historical comment information based on the recognition results of each of the multiple historical comment information; The sentiment analysis result determination submodule is used to input the text information contained in each of the at least one target comment into the sentiment analysis model to obtain the sentiment analysis result of the target object; and The profile information determination submodule is used to obtain the profile information of the target object based on the sentiment analysis results.

27. The apparatus according to claim 15, wherein, The target historical video determination module includes: The historical video determination submodule is used to determine, based on the video browsing history, multiple historical videos viewed by the target object within a preset time period; The log recording determination submodule is used to determine, based on the behavior logs, the log records corresponding to each of the multiple historical videos; and The target historical video determination submodule is used to determine at least one target historical video from the plurality of historical videos based on the log records corresponding to each of the plurality of historical videos.

28. The apparatus of claim 15, further comprising: The display module is used to display multiple target animated images included in the animated image resource set in a preset area in response to a comment input request.

29. An electronic device comprising: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-14.

30. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-14.

31. A computer program product comprising a computer program that, when executed by a processor, implements the method according to any one of claims 1-14.

Images