Video processing method and device, computer equipment and computer readable storage medium

Abstract

The embodiment of the invention discloses a video processing method and device, computer equipment and a computer readable storage medium, and the method comprises the steps: obtaining video titles of candidate videos and related text information of video contents, carrying out the entity extraction of the video titles and the related text information, and obtaining an entity word corresponding to each candidate video, the candidate videos are videos published by the same target user on the content platform; classifying the candidate videos according to the entity words to obtain candidate video sets belonging to different video series; for the candidate video set in each video series, calculating the similarity of every two candidate videos in the candidate video set in at least two dimensions; and screening the candidate videos in each candidate video set according to the similarity of every two candidate videos in each candidate video set in at least two dimensions to obtain a target video set of the target user in each video series. According to the embodiment of the invention, the candidate videos can be accurately classified, the video set of each video series is obtained, and the efficiency of searching the videos of the same video series is improved.

Description

Technical Field

[0001] The present application relates to the field of communication technology, and in particular to a video processing method, apparatus, computer equipment and computer-readable storage medium. Background Art

[0002] With the rapid development of Internet technology and the popularity of social media, in addition to sharing their own life moments on social media, people can also make some popular science or explanation videos based on their own interests and skills. For example, some users will publish series of videos, such as TV series or movie explanation video series, character introduction video series, and video series recording daily life, etc. Multiple videos produced on a specific theme. Since videos are usually sorted by release time, it is difficult for browsing users to quickly find videos of the same series. Summary of the invention

[0003] The embodiments of the present application provide a video processing method, apparatus, computer device, and computer-readable storage medium, which can accurately classify candidate videos and improve the efficiency of searching for similar videos.

[0004] A video processing method provided in an embodiment of the present application includes:

[0005] Obtaining the video title of the candidate video and the relevant text information of the video content, performing entity extraction on the video title and the relevant text information to obtain the entity words corresponding to each candidate video, wherein the candidate videos are videos released by the same target user on the content platform;

[0006] Classifying the candidate videos according to the entity words to obtain a collection of candidate videos belonging to different video series;

[0007] For each candidate video collection in each video series, calculating the similarity between every two candidate videos in the candidate video collection in at least two dimensions;

[0008] According to the similarity between every two candidate videos in each candidate video collection in at least two dimensions, the candidate videos in each candidate video collection are screened to obtain the target video collection of the target user under each video series.

[0009] Accordingly, an embodiment of the present application further provides a video processing device, including:

[0010] An acquisition unit is used to acquire a video title of a candidate video and relevant text information of the video content, and perform entity extraction on the video title and the relevant text information to obtain an entity word corresponding to each candidate video, wherein the candidate videos are videos released by the same target user on the content platform;

[0011] A classification unit, used to classify the candidate videos according to the entity words to obtain a collection of candidate videos belonging to different video series;

[0012] A calculation unit, configured to calculate, for each candidate video collection in each video series, the similarity between every two candidate videos in the candidate video collection in at least two dimensions;

[0013] The screening unit is used to screen the candidate videos in each candidate video collection according to the similarity between every two candidate videos in each candidate video collection in at least two dimensions, so as to obtain the target video collection of the target user under each video series.

[0014] In one embodiment, the computing unit comprises:

[0015] A first calculation subunit, configured to calculate, for each candidate video collection, a first similarity between every two candidate videos according to the video titles;

[0016] A second calculation subunit is used to calculate, for each candidate video collection, a second similarity between every two candidate videos according to the relevant text information;

[0017] The third calculation subunit is used to calculate, for each candidate video collection, a third similarity between every two candidate videos according to video frames of the candidate videos in the candidate video collection.

[0018] In one embodiment, the relevant text information includes content description text information and text recognition information, the second similarity includes description text similarity and recognition text similarity, and the second calculation subunit includes:

[0019] A first text similarity calculation module is used to calculate the description text similarity between every two candidate videos according to the content description text information for each candidate video collection;

[0020] The second text similarity calculation module is used to calculate the recognition text similarity between every two candidate videos according to the text recognition information for each candidate video collection.

[0021] In one embodiment, the third computing subunit includes:

[0022] An acquisition module, used to acquire a video cover image of each candidate video;

[0023] The first picture similarity calculation module is used to calculate the third similarity between every two candidate videos according to the video cover pictures for each candidate video collection.

[0024] In one embodiment, the screening unit comprises:

[0025] A clustering subunit, configured to cluster the candidate videos in each candidate video collection according to the similarity between every two candidate videos in each candidate video collection in at least two dimensions, to obtain at least one cluster;

[0026] The first collection determination subunit is used to determine the target video collection corresponding to each video series based on the number of candidate videos in the clustering cluster under each video series.

[0027] In one embodiment, the screening unit comprises:

[0028] A construction subunit is used to construct an undirected graph according to the similarity between every two candidate videos in each candidate video collection, wherein a vertex in the undirected graph corresponds to a candidate video, and the similarity of two candidate videos corresponding to two vertices connected by an edge in at least two dimensions meets a preset condition;

[0029] A maximal clique determining subunit, used for determining a maximal clique in the undirected graph;

[0030] The second collection determination subunit is used to determine the target video collection based on the candidate videos corresponding to the maximal clique.

[0031] In one embodiment, the screening unit comprises:

[0032] A mode acquisition subunit, used to acquire a play mode of each candidate video, wherein the play mode includes one of horizontal play and vertical play;

[0033] The third collection determination subunit is used to determine, for each candidate video collection, a target video in which a similar video exists in the candidate video collection, wherein the similarity between the similar video and the target video in at least one dimension is greater than a preset threshold and the playback modes are the same; and obtain the target video collection corresponding to the candidate video collection according to the target video.

[0034] In one embodiment, the classification unit comprises:

[0035] The vocabulary subunit is used to obtain a preset film and television entity vocabulary;

[0036] The matching subunit is used to match the entity word with the preset film and television entity word library to obtain the target film and television entity word and the candidate video collection under each target film and television entity word.

[0037] Correspondingly, an embodiment of the present application also provides a computer device, including a memory and a processor; the memory stores a computer program, and the processor is used to run the computer program in the memory to execute any video processing method provided in the embodiment of the present application.

[0038] Correspondingly, an embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium is used to store a computer program, and the computer program is loaded by a processor to execute any one of the video processing methods provided in the embodiments of the present application.

[0039] The embodiment of the present application obtains the video title of the candidate video and the relevant text information of the video content, performs entity extraction on the video title and the relevant text information, and obtains the entity words corresponding to each candidate video, wherein the candidate videos are videos released by the same target user on the content platform; classifies the candidate videos according to the entity words to obtain candidate video collections belonging to different video series; for the candidate video collection under each video series, calculates the similarity between every two candidate videos in the candidate video collection in at least two dimensions; based on the similarity between every two candidate videos in each candidate video collection in at least two dimensions, screens the candidate videos in each candidate video collection to obtain the target video collection of the target user under each video series.

[0040] The embodiment of the present application classifies the entity words corresponding to the candidate videos, and can obtain a candidate video collection corresponding to the entity words as the video series. The candidate video collection is then screened according to the similarity of the candidate videos in the candidate video collection in at least one dimension, and a video collection under the entity words as the video series can be obtained. The candidate videos can be accurately classified, thereby improving the efficiency of finding videos belonging to the same video series. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0042] Figure 1 is a flow chart of a video processing method provided by an embodiment of the present application;

[0043] Figure 2 is a schematic diagram of a video playback mode provided in an embodiment of the present application;

[0044] Figure 3 is a schematic diagram of video screening provided by an embodiment of the present application;

[0045] Figure 4 is a schematic diagram of a video processing device provided in an embodiment of the present application;

[0046] Figure 5 It is a schematic diagram of the structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

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

[0048] The embodiment of the present application provides a video processing method, apparatus, computer equipment and computer readable storage medium. The video processing apparatus can be integrated in a computer equipment, which can be a server or a terminal.

[0049] The terminal may include a mobile phone, a wearable smart device, a tablet computer, a laptop computer, a personal computer (PC), and a vehicle-mounted computer.

[0050] Among them, the server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, as well as big data and artificial intelligence platforms.

[0051] It should be noted that the description order of the following embodiments is not intended to limit the preferred order of the embodiments.

[0052] This embodiment will be described from the perspective of a video processing device. The video processing device may be integrated into a computer device, which may be a server or a terminal.

[0053] The embodiment of the present application provides a video processing method, such as Figure 1 As shown, the specific process of the video processing method can be as follows:

[0054] 101. Obtain video titles of candidate videos and relevant text information of video content, perform entity extraction on the video titles and the relevant text information, and obtain entity words corresponding to each candidate video, wherein the candidate videos are videos released by the same target user on the content platform.

[0055] Among them, the relevant text information can be information related to the video content of the candidate video, for example, it can include text information obtained by text recognition of the candidate video, and it can also include the introduction content entered by the user when publishing the candidate video, which can also be called the main text. When publishing a video, the user can add a video title and video text to explain the content of the published video.

[0056] For example, entity extraction may be performed on the video title and related text information of each candidate video respectively to obtain entity words contained in the video title and entity words contained in the related text information, thereby obtaining entity words corresponding to each candidate video.

[0057] Some users will publish a series of videos, such as a series of commentary videos of TV series or movies, a series of character introduction videos, and a series of videos recording daily life, etc. The entity words corresponding to the candidate videos may include the topics of the candidate videos, such as the names of TV series, movies, and characters.

[0058] The user can be used as the processing granularity, and the videos posted by each user (i.e., candidate videos) can be processed through steps 101 to 104. A collection can be constructed for the videos posted by each user to obtain a video collection corresponding to each keyword at the user granularity, so as to classify videos of different video series and obtain a video collection corresponding to each video series.

[0059] 102. Classify the candidate videos according to the entity words to obtain a collection of candidate videos belonging to different video series.

[0060] By classifying the candidate videos according to the entity words, a collection of candidate videos whose entity words belong to different categories can be obtained, and the entity words in each candidate video collection are the same or similar.

[0061] Optionally, a word library may be pre-set for matching with entity words of candidate videos. That is, in one embodiment, the step of “classifying the candidate videos according to the entity words to obtain a collection of candidate videos belonging to different video series” may specifically include:

[0062] Get the preset film and television entity vocabulary;

[0063] The entity words are matched with the preset film and television entity word library to obtain target film and television entity words and a collection of candidate videos under each target film and television entity word.

[0064] Among them, the preset film and television entity vocabulary may include TV series names and movie names, etc.

[0065] By matching the entity words of each candidate video with the entity words in the preset film and television vocabulary, a classification result using the entity words in the preset film and television vocabulary as keywords can be obtained.

[0066] Assuming that the preset film and television vocabulary includes three entity words A, B and C, the entity word of candidate video d is A, and the entity word of candidate video e is B, the candidate video d is matched with the preset film and television vocabulary to obtain the target film and television entity word A, and determine that the candidate video d is the video under the target film and television entity word A. Each candidate video is matched to obtain a candidate video collection of video series classified by the entity words in the preset film and television vocabulary.

[0067] It can be understood that each candidate video may correspond to one entity word, or more than one entity word. When there are multiple entity words, a candidate video may belong to at least one candidate video collection.

[0068] 103. For each candidate video collection in each video series, calculate the similarity between every two candidate videos in the candidate video collection in at least two dimensions.

[0069] Among them, at least two dimensions may include a video title, text information related to the video content, and at least two dimensions of a video frame and a video cover image.

[0070] For example, for each video series, the similarity between every two candidate videos in at least two dimensions may be calculated, and whether the two candidate videos belong to the same video collection may be determined based on the similarity between the two candidate videos in at least two dimensions.

[0071] In one embodiment, at least two dimensions may include video titles and related text information. The similarity between the video titles of every two candidate videos and the similarity between the related text information of every two videos may be calculated. If the similarity in one dimension is greater than a preset threshold, the two candidate videos are considered to belong to the same video collection.

[0072] Optionally, it is also possible to determine whether two candidate videos belong to the same video collection in combination with the video content, that is, at least two dimensions may also include video frames. In one embodiment, the step of "for each candidate video collection under each video series, calculating the similarity between every two candidate videos in the candidate video collection in at least two dimensions" may specifically include:

[0073] For each candidate video collection, calculating a first similarity between every two candidate videos according to the video titles;

[0074] For each candidate video collection, calculating a second similarity between every two candidate videos according to the relevant text information;

[0075] For each candidate video collection, the third similarity between every two candidate videos is calculated according to the video frames of the candidate videos in the candidate video collection.

[0076] For example, specifically, for each candidate video collection, the similarity between the video titles of every two candidate videos therein is calculated to obtain a first similarity; the second similarity between the relevant text information of every two candidate videos therein is calculated; the similarity between the video frames of every two candidate videos therein is calculated, which may be by performing feature extraction on each candidate video to obtain video feature information of each candidate video, and then determining the similarity between the two candidate videos based on the distance between the video feature information of the two candidate videos in the feature space, wherein the greater the distance, the smaller the similarity.

[0077] In one embodiment, the relevant text information may include content description text information and text recognition information, the text recognition information may include text information obtained by performing text recognition on the candidate video, the content description information may include the introduction content entered by the user when publishing the candidate video, which may also be called the main text, so the second similarity may include the similarity between the content description texts and the similarity between the text recognition information, that is, in one embodiment, the relevant text information includes content description text information and text recognition information, the second similarity includes description text similarity and recognition text similarity, and the step of "for each candidate video collection, calculating the second similarity between each two candidate videos according to the relevant text information" includes:

[0078] For each candidate video collection, calculating the description text similarity between every two candidate videos according to the content description text information;

[0079] For each candidate video collection, the recognition text similarity between every two candidate videos is calculated according to the text recognition information.

[0080] For example, for each candidate video collection, the description text similarity between every two candidate videos is calculated based on the content description text information; for each candidate video collection, the recognition text similarity between every two candidate videos is calculated based on the text recognition information.

[0081] The third similarity may be the similarity between video frames of the candidate videos. For example, feature extraction may be performed on the candidate videos to obtain video feature information of the candidate videos, and then the similarity between two candidate videos is determined based on the video feature information of every two candidate videos.

[0082] Optionally, the third similarity may also be the similarity between video cover images. That is, in one embodiment, the step of “for each candidate video collection, calculating the third similarity between every two candidate videos according to the video frames of the candidate videos in the candidate video collection” may specifically include:

[0083] Get the video cover image of each candidate video;

[0084] For each candidate video collection, the third similarity between every two candidate videos is calculated based on the video cover images.

[0085] The video cover image can be the cover image that the user sees when the video is not played, or it can be the first frame of the video played.

[0086] For example, the video cover image of each candidate video may be obtained, and then for each candidate video collection, the similarity between the video cover images of each two candidate videos may be calculated to obtain the third similarity of the candidate videos.

[0087] 104. According to the similarity between every two candidate videos in each candidate video collection in at least two dimensions, the candidate videos in each candidate video collection are screened to obtain a target video collection for the target user in each video series.

[0088] In step 104, the similarity between any two candidate videos in each candidate video collection in at least two dimensions is obtained. Based on the similarity between the two candidate videos in at least two dimensions, it can be determined whether the two candidate videos belong to the same video collection. Based on this, candidate videos belonging to the same video collection can be screened from each candidate video collection to obtain the target video collection.

[0089] In one embodiment, the candidate videos belonging to the same collection may be determined in combination with the similarity and the video playback mode. That is, in one embodiment, the step of "screening the candidate videos in each candidate video collection according to the similarity between every two candidate videos in each candidate video collection in at least two dimensions to obtain a target video collection corresponding to each video series" includes:

[0090] Obtaining a playback mode of each candidate video, where the playback mode includes one of horizontal playback and vertical playback;

[0091] For each candidate video collection, determining a target video having a similar video in the candidate video collection, wherein the similarity between the similar video and the target video in at least one dimension is greater than a preset threshold and the play modes are the same;

[0092] A target video collection corresponding to the candidate video collection is obtained according to the target video.

[0093] The schematic diagram of horizontal version playback and vertical version playback can be as follows: Figure 2 As shown, if the similarity between two candidate videos in at least one dimension is greater than a preset threshold, and the information of the candidate videos in the dimension is not empty, and the playback mode is the same, then the two candidate videos are considered similar and belong to the same target video collection.

[0094] For example, you can Figure 3As shown, if the candidate video has a video title, that is, the video title is not empty, and the similarity of the candidate titles is greater than a preset threshold and the playback mode is the same, then they belong to the same video collection; if the candidate video has a text, and the similarity of the text is greater than a preset threshold and the playback mode is the same, then they belong to the same video collection.

[0095] The candidate video collection can be represented as a tree structure (which can include at least one tree) based on the candidate videos in the candidate video collection and whether two candidate videos belong to the same target video collection. In the tree structure, each node corresponds to a candidate video, and two similar candidate videos are connected by edges. Each tree can correspond to a target video collection.

[0096] A depth threshold may also be set, and each tree may be traversed using a depth-first traversal algorithm to obtain at least one target video collection.

[0097] A target video collection may be selected from at least one target video collection as the final target video collection, for example, a target video collection containing the largest number of videos may be selected. Alternatively, in one embodiment, the step of "screening candidate videos in each candidate video collection according to the similarity between every two candidate videos in each candidate video collection in at least two dimensions to obtain a target video collection corresponding to each video series" may include:

[0098] Clustering the candidate videos in each candidate video collection according to the similarity between every two candidate videos in at least two dimensions in each candidate video collection to obtain at least one cluster;

[0099] Based on the number of candidate videos in the clusters under each video series, the target video collection corresponding to each video series is determined.

[0100] For example, when the similarity between two candidate videos in at least one dimension is greater than a preset threshold, the two candidate videos can be considered to belong to the same video collection, and the candidate videos belonging to the same video collection are clustered. Optionally, clustering can also be performed in combination with the playback mode of the candidate videos to cluster the candidate videos in each candidate video collection into at least one clustering cluster.

[0101] According to the number of candidate videos contained in the clustering clusters, the video clusters are sorted, and the video cluster containing the largest number of candidate videos is selected. The target video collection is obtained based on the candidate videos contained in the video cluster. Alternatively, a clustering cluster containing candidate video data greater than 80% of the number of candidate videos in the candidate video collection can be selected.

[0102] Optionally, an undirected graph may be constructed according to the similarity between two candidate videos in at least two dimensions, and the candidate video collection may be screened according to the undirected graph. That is, in one embodiment, the step of "screening the candidate videos in each candidate video collection according to the similarity between every two candidate videos in each candidate video collection in at least two dimensions to obtain a target video collection corresponding to each video series" may include:

[0103] Constructing an undirected graph according to the similarity between every two candidate videos in each candidate video collection, wherein a vertex in the undirected graph corresponds to a candidate video, and the similarity between two candidate videos corresponding to two vertices connected by an edge in at least two dimensions meets a preset condition;

[0104] determining a maximal clique in the undirected graph;

[0105] The target video collection is determined according to the candidate videos corresponding to the maximal clique.

[0106] A clique is a complete subgraph of an undirected graph, and a complete subgraph is a subgraph in which every pair of vertices is connected by an edge. If a clique is not contained in any other clique, that is, it is not a proper subset of any other clique, then the clique is called a maximal clique of an undirected graph. The maximal clique is an independent complete subset of an undirected graph.

[0107] For example, specifically, for each candidate video collection, an undirected graph may be constructed according to the similarity between every two candidate videos in the candidate video collection, where each vertex in the undirected graph represents a candidate video, and if the similarity between two candidate videos in at least one dimension is greater than a preset threshold, the vertices corresponding to the two candidate videos are connected by edges.

[0108] Determine the maximal clique of the undirected graph corresponding to each candidate video collection, and obtain the target video collection under each video series based on the candidate videos included in the maximal clique. The video series can be based on entity words in a preset film and television vocabulary as keywords, so the target video collection under each keyword can be obtained.

[0109] As can be seen from the above, the embodiment of the present application obtains the video title of the candidate video and the relevant text information of the video content, performs entity extraction on the video title and the relevant text information, and obtains the entity words corresponding to each candidate video, wherein the candidate video is a video released by the same target user on the content platform; the candidate videos are classified according to the entity words to obtain candidate video collections belonging to different video series; for the candidate video collection under each video series, the similarity between every two candidate videos in the candidate video collection in at least two dimensions is calculated; according to the similarity between every two candidate videos in each candidate video collection in at least two dimensions, the candidate videos in each candidate video collection are screened to obtain the target video collection of the target user under each video series.

[0110] The embodiment of the present application classifies the entity words corresponding to the candidate videos, and can obtain a candidate video collection corresponding to the entity words as the video series. Then, the candidate video collection is screened according to the similarity of the candidate videos in at least one dimension, and a video collection under the entity words as the video series can be obtained. The candidate videos can be accurately classified, and the efficiency of finding videos belonging to the same video series is improved. .

[0111] In order to better implement the video processing method provided in the embodiment of the present application, a video processing device is also provided in one embodiment. The meanings of the terms are the same as those in the above-mentioned video processing method, and the specific implementation details can refer to the description in the method embodiment.

[0112] The video processing device can be integrated into a computer device, such as Figure 4 As shown, the video processing device may include: an acquisition unit 301, a classification unit 302, a calculation unit 303 and a screening unit 304, which are specifically as follows:

[0113] (1) An acquisition unit 301 is used to acquire a video title of a candidate video and relevant text information of the video content, perform entity extraction on the video title and relevant text information, and obtain entity words corresponding to each candidate video, wherein the candidate videos are videos released by the same target user on the content platform.

[0114] (2) A classification unit 302 is used to classify candidate videos according to entity words to obtain a collection of candidate videos belonging to different video series.

[0115] In one embodiment, the classification unit 302 includes:

[0116] The vocabulary subunit is used to obtain a preset film and television entity vocabulary;

[0117] The matching subunit is used to match the entity words with the preset film and television entity word library to obtain the target film and television entity words and the candidate video collection under each target film and television entity word.

[0118] (3) A calculation unit 303 is used to calculate the similarity between every two candidate videos in the candidate video collection in at least two dimensions for each candidate video collection in each video series.

[0119] In one embodiment, the calculation unit 303 includes:

[0120] A first calculation subunit, configured to calculate, for each candidate video collection, a first similarity between every two candidate videos according to the video titles;

[0121] A second calculation subunit is used to calculate the second similarity between every two candidate videos according to the relevant text information for each candidate video collection;

[0122] The third calculation subunit is used to calculate, for each candidate video collection, a third similarity between every two candidate videos according to video frames of the candidate videos in the candidate video collection.

[0123] In one embodiment, the relevant text information includes content description text information and text recognition information, the second similarity includes description text similarity and recognition text similarity, and the second calculation subunit includes:

[0124] A first text similarity calculation module is used to calculate the description text similarity between every two candidate videos according to the content description text information for each candidate video collection;

[0125] The second text similarity calculation module is used to calculate the recognition text similarity between every two candidate videos according to the text recognition information for each candidate video collection.

[0126] In one embodiment, the third computing subunit includes:

[0127] An acquisition module, used to acquire a video cover image of each candidate video;

[0128] The first picture similarity calculation module is used to calculate the third similarity between every two candidate videos according to the video cover pictures for each candidate video collection.

[0129] (4) A screening unit 304 is used to screen the candidate videos in each candidate video collection according to the similarity between every two candidate videos in each candidate video collection in at least two dimensions, so as to obtain a target video collection for the target user in each video series.

[0130] In one embodiment, the screening unit 304 includes:

[0131] A clustering subunit, configured to cluster the candidate videos in each candidate video collection according to the similarity between every two candidate videos in each candidate video collection in at least two dimensions, to obtain at least one cluster;

[0132] The first collection determination subunit is used to determine the target video collection corresponding to each video series based on the number of candidate videos in the clustering cluster under each video series.

[0133] In one embodiment, the screening unit 304 includes:

[0134] A mode acquisition subunit is used to acquire a playback mode of each candidate video, where the playback mode includes one of horizontal playback and vertical playback;

[0135] The third collection determination subunit is used to determine, for each candidate video collection, a target video in which a similar video exists in the candidate video collection, wherein the similarity between the similar video and the target video in at least one dimension is greater than a preset threshold and the playback modes are the same; and obtain a target video collection corresponding to the candidate video collection according to the target video.

[0136] In one embodiment, the screening unit 304 includes:

[0137] A construction subunit is used to construct an undirected graph according to the similarity between every two candidate videos in each candidate video collection, wherein a vertex in the undirected graph corresponds to a candidate video, and the similarity between two candidate videos corresponding to two vertices connected by an edge in at least two dimensions meets a preset condition;

[0138] The maximum clique determination subunit is used to determine the maximum clique in an undirected graph;

[0139] The second collection determination subunit is used to determine the target video collection based on the candidate videos corresponding to the maximum clique.

[0140] As can be seen from the above, the video processing device of the embodiment of the present application obtains the video title of the candidate video and the relevant text information of the video content, performs entity extraction on the video title and the relevant text information, and obtains the entity words corresponding to each candidate video, wherein the candidate video is a video released by the same target user on the content platform; classifies the candidate videos according to the entity words to obtain candidate video collections belonging to different video series; for the candidate video collection under each video series, calculates the similarity between every two candidate videos in the candidate video collection in at least two dimensions; according to the similarity between every two candidate videos in each candidate video collection in at least two dimensions, screens the candidate videos in each candidate video collection to obtain the target video collection corresponding to each video series.

[0141] The embodiment of the present application classifies the entity words corresponding to the candidate videos, and can obtain a candidate video collection corresponding to the entity words as the video series. Then, the candidate video collection is screened according to the similarity of the candidate videos in at least one dimension, and a video collection under the entity words as the video series can be obtained. The candidate videos can be accurately classified, and the efficiency of finding videos belonging to the same video series is improved. .

[0142] The present application also provides a computer device, which may be a terminal or a server. Figure 5 As shown, it shows a schematic diagram of the structure of the computer device involved in the embodiment of the present application, specifically:

[0143] The computer device may include components such as a processor 1001 with one or more processing cores, a memory 1002 with one or more computer-readable storage media, a power supply 1003, and an input unit 1004. Those skilled in the art will appreciate that Figure 5 The computer device structure shown in the figure does not constitute a limitation on the computer device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently. Among them:

[0144] The processor 1001 is the control center of the computer device. It uses various interfaces and lines to connect various parts of the entire computer device. By running or executing software programs and / or modules stored in the memory 1002 and calling data stored in the memory 1002, it executes various functions of the computer device and processes data, thereby monitoring the computer device as a whole. Optionally, the processor 1001 may include one or more processing cores; preferably, the processor 1001 may integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface and computer programs, etc., and the modem processor mainly processes wireless communications. It is understandable that the above-mentioned modem processor may not be integrated into the processor 1001.

[0145] The memory 1002 can be used to store software programs and modules. The processor 1001 executes various functional applications and data processing by running the software programs and modules stored in the memory 1002. The memory 1002 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, a computer program required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 1002 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage devices. Accordingly, the memory 1002 may also include a memory controller to provide the processor 1001 with access to the memory 1002.

[0146] The computer device also includes a power supply 1003 for supplying power to various components. Preferably, the power supply 1003 can be logically connected to the processor 1001 through a power management system, so as to manage charging, discharging, and power consumption through the power management system. The power supply 1003 can also include any components such as one or more DC or AC power supplies, recharging systems, power failure detection circuits, power converters or inverters, and power status indicators.

[0147] The computer device may further include an input unit 1004, which may be used to receive input digital or character information and generate keyboard, mouse, joystick, optical or trackball signal input related to user settings and function control.

[0148] Although not shown, the computer device may further include a display unit, etc., which will not be described in detail herein. Specifically, in this embodiment, the processor 1001 in the computer device will load the executable files corresponding to the processes of one or more computer programs into the memory 1002 according to the following instructions, and the processor 1001 will run the computer programs stored in the memory 1002, thereby realizing various functions, as follows:

[0149] Obtain the video title of the candidate video and the relevant text information of the video content, perform entity extraction on the video title and the relevant text information, and obtain the entity words corresponding to each candidate video, wherein the candidate videos are videos released by the same target user on the content platform;

[0150] Classify candidate videos according to entity words to obtain a collection of candidate videos belonging to different video series;

[0151] For each candidate video collection under each video series, calculate the similarity between every two candidate videos in the candidate video collection in at least two dimensions;

[0152] According to the similarity between every two candidate videos in each candidate video collection in at least two dimensions, the candidate videos in each candidate video collection are screened to obtain a target video collection for the target user under each video series.

[0153] As can be seen from the above, the computer device of the embodiment of the present application can obtain the video title of the candidate video and the relevant text information of the video content, perform entity extraction on the video title and the relevant text information, and obtain the entity words corresponding to each candidate video, wherein the candidate video is a video released by the same target user on the content platform; classify the candidate videos according to the entity words to obtain candidate video collections belonging to different video series; for the candidate video collection under each video series, calculate the similarity between every two candidate videos in the candidate video collection in at least two dimensions; according to the similarity between every two candidate videos in each candidate video collection in at least two dimensions, screen the candidate videos in each candidate video collection to obtain the target video collection of the target user under each video series.

[0154] The embodiment of the present application classifies the entity words corresponding to the candidate videos, and can obtain a candidate video collection corresponding to the entity words as the video series. The candidate video collection is then screened according to the similarity of the candidate videos in the candidate video collection in at least one dimension, and a video collection under the entity words as the video series can be obtained. The candidate videos can be accurately classified, thereby improving the efficiency of finding videos belonging to the same video series.

[0155] According to one aspect of the present application, a computer program product or a computer program is provided, the computer program product or the computer program including computer instructions, the computer instructions being stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the methods provided in various optional implementations in the above-mentioned embodiments.

[0156] Those skilled in the art will appreciate that all or part of the steps in the various methods of the above embodiments may be completed by a computer program, or by controlling related hardware through a computer program. The computer program may be stored in a computer-readable storage medium and loaded and executed by a processor.

[0157] To this end, an embodiment of the present application provides a computer-readable storage medium, in which a computer program is stored. The computer program can be loaded by a processor to execute any video processing method provided in the embodiment of the present application.

[0158] The specific implementation of the above operations can be found in the previous embodiments, which will not be described in detail here.

[0159] The computer-readable storage medium may include: a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0160] Since the computer program stored in the computer-readable storage medium can execute any one of the video processing methods provided in the embodiments of the present application, the beneficial effects that can be achieved by any one of the video processing methods provided in the embodiments of the present application can be achieved. Please refer to the previous embodiments for details and will not be repeated here.

[0161] The above is a detailed introduction to a video processing method, device, computer equipment and computer-readable storage medium provided in the embodiments of the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for technical personnel in this field, according to the ideas of the present application, there will be changes in the specific implementation methods and application scopes. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A video processing method, characterized in that: include: Obtaining the video title of the candidate video and the relevant text information of the video content, performing entity extraction on the video title and the relevant text information to obtain the entity words corresponding to each candidate video, wherein the candidate videos are videos released by the same target user on the content platform; Classifying the candidate videos according to the entity words to obtain a collection of candidate videos belonging to different video series; For each candidate video collection in each video series, calculating the similarity between every two candidate videos in the candidate video collection in at least two dimensions; According to the similarity between every two candidate videos in each candidate video collection in at least two dimensions, the candidate videos in each candidate video collection are screened to obtain the target video collection of the target user under each video series.

2. The method according to claim 1, characterized in that The step of calculating the similarity between every two candidate videos in the candidate video collection in at least two dimensions for each candidate video collection in each video series includes: For each candidate video collection, calculating a first similarity between every two candidate videos according to the video titles; For each candidate video collection, calculating a second similarity between every two candidate videos according to the relevant text information; For each candidate video collection, the third similarity between every two candidate videos is calculated according to the video frames of the candidate videos in the candidate video collection.

3. The method according to claim 2, characterized in that The relevant text information includes content description text information and text recognition information, the second similarity includes description text similarity and recognition text similarity, and for each candidate video collection, calculating the second similarity of every two candidate videos according to the relevant text information includes: For each candidate video collection, calculating the description text similarity between every two candidate videos according to the content description text information; For each candidate video collection, the recognition text similarity between every two candidate videos is calculated according to the text recognition information.

4. The method according to claim 2, characterized in that: The step of calculating, for each candidate video collection, a third similarity between every two candidate videos according to video frames of the candidate videos in the candidate video collection includes: Get the video cover image of each candidate video; For each candidate video collection, the third similarity between every two candidate videos is calculated based on the video cover images.

5. The method according to claim 1, characterized in that The method of screening the candidate videos in each candidate video collection according to the similarity between every two candidate videos in at least two dimensions to obtain a target video collection corresponding to each video series includes: Obtaining a playback mode of each candidate video, where the playback mode includes one of horizontal playback and vertical playback; For each candidate video collection, determining a target video having a similar video in the candidate video collection, wherein the similarity between the similar video and the target video in at least one dimension is greater than a preset threshold and the play modes are the same; A target video collection corresponding to the candidate video collection is obtained according to the target video.

6. The method according to claim 1, characterized in that The method of screening the candidate videos in each candidate video collection according to the similarity between every two candidate videos in at least two dimensions to obtain a target video collection corresponding to each video series includes: Clustering the candidate videos in each candidate video collection according to the similarity between every two candidate videos in at least two dimensions in each candidate video collection to obtain at least one cluster; Based on the number of candidate videos in the clusters under each video series, the target video collection corresponding to each video series is determined.

7. The method according to claim 1, characterized in that The method of screening the candidate videos in each candidate video collection according to the similarity between every two candidate videos in at least two dimensions to obtain a target video collection corresponding to each video series includes: Constructing an undirected graph according to the similarity between every two candidate videos in each candidate video collection, wherein a vertex in the undirected graph corresponds to a candidate video, and the similarity between two candidate videos corresponding to two vertices connected by an edge in at least two dimensions meets a preset condition; determining a maximal clique in the undirected graph; The target video collection is determined according to the candidate videos corresponding to the maximal clique.

8. The method according to claim 1, characterized in that The classifying the candidate videos according to the entity words to obtain a collection of candidate videos belonging to different video series includes: Get the preset film and television entity vocabulary; The entity words are matched with the preset film and television entity word library to obtain target film and television entity words and a collection of candidate videos under each target film and television entity word.

9. A video processing device, characterized in that: include: An acquisition unit is used to acquire a video title of a candidate video and relevant text information of the video content, and perform entity extraction on the video title and the relevant text information to obtain an entity word corresponding to each candidate video, wherein the candidate videos are videos released by the same target user on the content platform; A classification unit, used to classify the candidate videos according to the entity words to obtain a collection of candidate videos belonging to different video series; A calculation unit, configured to calculate, for each candidate video collection in each video series, the similarity between every two candidate videos in the candidate video collection in at least two dimensions; The screening unit is used to screen the candidate videos in each candidate video collection according to the similarity between every two candidate videos in each candidate video collection in at least two dimensions, so as to obtain the target video collection of the target user under each video series.

10. A computer device, characterized in that: It comprises a memory and a processor; the memory stores a computer program, and the processor is used to run the computer program in the memory to execute the video processing method according to any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium is used to store a computer program, and the computer program is loaded by a processor to execute the video processing method according to any one of claims 1 to 8.

Images