Video analysis method and device, electronic device and storage medium

Abstract

The present application relates to a video analysis method and device, an electronic device, and a storage medium. The method includes: obtaining a video to be analyzed; determining a first target evaluation and a second target evaluation corresponding to a target video segment according to the intra-video evaluation information and inter-video evaluation information corresponding to the video to be analyzed; determining a target wonderfulness evaluation value corresponding to the target video segment based on the intra-video evaluation information, the inter-video evaluation information, the first target evaluation, and the second target evaluation; determining a wonderfulness category corresponding to the target video segment according to the target wonderfulness evaluation value corresponding to the target video segment. The method provided by the present application can effectively overcome the technical problems in the related art of manually labeling the wonderfulness of videos, which are highly subjective, slow, and unable to achieve rapid batch data production.

Description

Technical Field

[0001] The present application relates to the field of video analysis technology, and in particular to a video analysis method and device, an electronic device, and a storage medium. Background Art

[0002] When developing a video highlight analysis algorithm, a large amount of training data is required for model training. However, the highlight level of a video is highly subjective, especially when the highlight level of a clip is represented by a score from 0 to 10 rather than a label. Manual data annotation makes it difficult to objectively and accurately assign a reasonable score. Furthermore, when adding to the training dataset, the data annotators may need to be replaced for various reasons, making it even more difficult to achieve uniform annotation standards. Furthermore, manual data annotation is slow, making it impossible to quickly and efficiently produce batches of data.

[0003] It can be seen from this that there are technical problems in the relevant technology that manual labeling of the excitement of videos is highly subjective, the labeling speed is slow, and it is impossible to achieve fast batch data production. Summary of the Invention

[0004] In order to solve the technical problems that manual labeling of video excitement is highly subjective, slow, and unable to achieve rapid batch data production, the present application provides a video analysis method and device, an electronic device, and a storage medium.

[0005] In a first aspect, an embodiment of the present application provides a video analysis method, comprising:

[0006] Get the video to be analyzed;

[0007] Determine a first target evaluation and a second target evaluation corresponding to a target video segment according to intra-video evaluation information and inter-video evaluation information corresponding to the video to be analyzed, wherein the intra-video evaluation information includes the first target evaluation corresponding to the target video segment, the inter-video evaluation information includes the second target evaluation corresponding to the target video segment, the first target evaluation is determined according to the evaluation of each candidate video segment in the video to be analyzed, the second target evaluation is a global evaluation determined according to the evaluation of each video segment in a target video set, all candidate video segments include the target video segment, each candidate video segment has a unique corresponding time period in the video to be analyzed, and the target video set includes the video to be analyzed;

[0008] Determining a target excitement evaluation value corresponding to the target video segment based on the intra-video evaluation information, the inter-video evaluation information, the first target evaluation, and the second target evaluation;

[0009] According to the target excitement evaluation value corresponding to the target video segment, an excitement category corresponding to the target video segment is determined.

[0010] Optionally, as in the aforementioned method, determining a target excitement evaluation value corresponding to the target video segment based on the intra-video evaluation information, the inter-video evaluation information, the first target evaluation, and the second target evaluation includes:

[0011] Determining, according to the in-video evaluation information, a first candidate evaluation corresponding to each candidate video segment;

[0012] Calculate the average value of all first candidate evaluations to obtain a first average value;

[0013] Determining, according to the inter-video evaluation information, a second candidate evaluation corresponding to each candidate video segment;

[0014] Calculate the average value of all second candidate evaluations to obtain a second average value;

[0015] Determining a first magnitude relationship between the first target evaluation and the first average value; determining a second magnitude relationship between the second target evaluation and the second average value;

[0016] A target excitement evaluation value corresponding to the target video segment is determined based on the first size relationship and the second size relationship.

[0017] Optionally, as in the aforementioned method, determining the first magnitude relationship between the first target evaluation and the first average value includes:

[0018] Determining a first minimum candidate evaluation with the lowest evaluation value and a first maximum candidate evaluation with the highest evaluation value among all first candidate evaluations;

[0019] Dividing the first minimum candidate evaluation and the first average value to obtain a first preset number of first low evaluation value intervals; dividing the first maximum candidate evaluation and the first average value to obtain a second preset number of first high evaluation value intervals;

[0020] By determining a first target evaluation value interval including the first target evaluation value in all first evaluation value intervals, a first size relationship between the first target evaluation and the first average value is obtained, wherein all first evaluation value intervals include the first low evaluation value interval and the first high evaluation value interval.

[0021] Optionally, as in the aforementioned method, determining the second magnitude relationship between the second target evaluation and the second average value includes:

[0022] Determining a second minimum candidate evaluation with the lowest evaluation value and a second maximum candidate evaluation with the highest evaluation value among all second candidate evaluations;

[0023] Dividing the second minimum candidate evaluation and the second average value to obtain a third preset number of second low evaluation value intervals; determining the minimum designated second evaluation between the second highest candidate evaluation and the preset second evaluation upper limit, and dividing the designated second evaluation and the second average value to obtain a fourth preset number of second high evaluation value intervals;

[0024] A second size relationship between the second target evaluation and the second average value is obtained by determining a second target evaluation value interval including the second target evaluation value in all second evaluation value intervals, wherein all second evaluation value intervals include the second low evaluation value interval and the second high evaluation value interval.

[0025] Optionally, as in the aforementioned method, determining the target excitement evaluation value corresponding to the target video segment based on the first size relationship and the second size relationship includes:

[0026] Determine a first wonderfulness value corresponding to each first evaluation value interval; determine a second wonderfulness value corresponding to each second evaluation value interval;

[0027] Determining a first target exciting value corresponding to the first target exciting value interval according to the first exciting value corresponding to each first evaluation value interval; determining a second target exciting value corresponding to the second target exciting value interval according to the second exciting value corresponding to each second evaluation value interval;

[0028] Obtaining the target excitement evaluation value by calculating the first target excitement value and the second target excitement value according to a preset weighting method;

[0029] When the second target evaluation value is higher than or equal to the preset second evaluation upper limit, the target video segment is assigned the maximum target excitement evaluation value.

[0030] Optionally, as in the aforementioned method, determining the excitement category corresponding to the target video segment according to the target excitement evaluation value corresponding to the target video segment includes:

[0031] Performing behavior detection on the target video clip to obtain a behavior detection result;

[0032] If the behavior detection result indicates that a behavior of a preset behavior type exists in the target video segment, and the target excitement evaluation value is greater than or equal to a preset threshold, determining that the excitement category corresponding to the target video segment is a high excitement category;

[0033] When the behavior detection result indicates that a behavior of a preset behavior type exists in the target video segment and the target excitement evaluation value is less than a preset threshold, the excitement category corresponding to the target video segment is determined to be a low excitement category.

[0034] Optionally, as in the aforementioned method, after determining the excitement category corresponding to the target video segment according to the target excitement evaluation value corresponding to the target video segment, the method further includes:

[0035] Determining a designated excitement category corresponding to each designated video segment, wherein all designated video segments include the target video segment;

[0036] The designated video segment corresponding to the designated excitement category being a high excitement category is used as a high excitement segment;

[0037] The designated video segment corresponding to the designated excitement category being a low excitement category is used as a low excitement segment;

[0038] According to a preset quantity relationship, a first number of high-highlight training segments are selected from all the high-highlight segments, and a second number of low-highlight training segments are selected from all the low-highlight segments, wherein the first number and the second number satisfy the preset quantity relationship.

[0039] In a second aspect, an embodiment of the present application provides a video analysis device, comprising:

[0040] An acquisition module is used to acquire the video to be analyzed;

[0041] A first determination module is configured to determine a first target evaluation and a second target evaluation corresponding to a target video segment according to intra-video evaluation information and inter-video evaluation information corresponding to the video to be analyzed, wherein the intra-video evaluation information includes the first target evaluation corresponding to the target video segment, the inter-video evaluation information includes the second target evaluation corresponding to the target video segment, the first target evaluation is determined according to the evaluation of each candidate video segment in the video to be analyzed, the second target evaluation is a global evaluation determined according to the evaluation of each video segment in a target video set, all candidate video segments include the target video segment, each candidate video segment has a unique corresponding time period in the video to be analyzed, and the target video set includes the video to be analyzed;

[0042] a second determining module, configured to determine a target excitement evaluation value corresponding to the target video segment based on the intra-video evaluation information, the inter-video evaluation information, the first target evaluation, and the second target evaluation;

[0043] The excitement category determination module is configured to determine an excitement category corresponding to the target video segment according to a target excitement evaluation value corresponding to the target video segment.

[0044] In a third aspect, an embodiment of the present application provides an electronic device, comprising: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus;

[0045] The memory is used to store computer programs;

[0046] The processor is configured to implement any of the aforementioned methods when executing the computer program.

[0047] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the storage medium includes a stored program, wherein the program executes the method as described in any of the preceding items when running.

[0048] The above technical solution provided by the embodiment of the present application has the following advantages compared with the prior art:

[0049] The method provided in the embodiment of the present application provides an implementation method for automatically determining the excitement category corresponding to a video clip. Compared with the manual labeling of the excitement of the video in the related art, it can effectively ensure the uniformity of the judgment standard and improve the efficiency of determining the excitement category, thereby effectively overcoming the technical problems of the manual labeling of the excitement of the video in the related art, which is highly subjective, slow, and unable to achieve fast batch data production. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0051] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0052] Figure 1 A schematic diagram of a video analysis method provided in an embodiment of the present application;

[0053] Figure 2 This is a schematic diagram of the filter scores and filter-related scores corresponding to Video 1 in the application example of this application;

[0054] Figure 3 This is a schematic diagram of the filter scores and filter-related scores corresponding to Video 2 in the application example of this application;

[0055] Figure 4 A block diagram of a video analysis device provided in an embodiment of the present application;

[0056] Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0057] To make the purpose, technical solutions, and advantages of the embodiments of this application more clear, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0058] According to one aspect of an embodiment of the present application, a video analysis method is provided. Optionally, in this embodiment, the video analysis method can be applied to a hardware environment consisting of a terminal and a server. The server is connected to the terminal via a network and can be used to provide services (such as advertising push services, application services, etc.) for the terminal or a client installed on the terminal. A database can be set up on the server or independently of the server to provide data storage services for the server.

[0059] The aforementioned network may include, but is not limited to, at least one of the following: a wired network and a wireless network. The aforementioned wired network may include, but is not limited to, at least one of the following: a wide area network, a metropolitan area network, or a local area network. The aforementioned wireless network may include, but is not limited to, at least one of the following: Wi-Fi (Wireless Fidelity) and Bluetooth. The terminal may be, but is not limited to, a PC, a mobile phone, a tablet computer, or the like.

[0060] The video analysis method of the embodiment of the present application can be executed by a server, a terminal, or both. The video analysis method of the embodiment of the present application can also be executed by a client installed on the terminal.

[0061] Taking the video analysis method of this embodiment executed by the server as an example, Figure 1 A video analysis method provided in an embodiment of the present application includes the following steps:

[0062] Step S101: Obtain the video to be analyzed.

[0063] The video analysis method of this embodiment can be applied to scenarios where it is necessary to identify exciting and non-exciting video clips in a video, for example, a scenario where a model of an exciting analysis algorithm is trained using exciting and non-exciting video clips.

[0064] Taking the analysis of a video to be analyzed as an example, one or more candidate video segments in the video to be analyzed are identified to determine the excitement of the one or more candidate video segments.

[0065] The video to be analyzed may be actively obtained by the server implementing the method of this embodiment from a video library, or may be uploaded to the server.

[0066] The video to be analyzed may include one or more videos, and the steps of the method of this embodiment may be executed sequentially or concurrently between different videos.

[0067] For example, the server obtains one of the videos (eg, a video of an episode of a TV series, a video of a movie, etc.) from the database of the video platform and uses it as the video to be analyzed.

[0068] Step S102: Determine a first target evaluation and a second target evaluation corresponding to the target video segment according to the in-video evaluation information and the inter-video evaluation information corresponding to the video to be analyzed, wherein the in-video evaluation information includes the first target evaluation corresponding to the target video segment, and the inter-video evaluation information includes the second target evaluation corresponding to the target video segment. The first target evaluation is determined according to the evaluation of each candidate video segment in the video to be analyzed, and the second target evaluation is a global evaluation determined according to the evaluation of each video segment in the target video set. The target video segment is included in all candidate video segments, and each candidate video segment has a unique corresponding time period in the video to be analyzed. The target video segment is included in the target video set.

[0069] After obtaining the video to be analyzed, intra-video evaluation information and inter-video evaluation information corresponding to the video to be analyzed may be obtained.

[0070] The intra-video evaluation information and the inter-video evaluation information may be information stored in a designated database in association with the video to be analyzed.

[0071] In the video to be analyzed, each candidate video segment has a unique corresponding time period (for example, 1 second, 2 seconds, etc.), and the time periods corresponding to different candidate video segments do not overlap, and generally, the candidate video segments are continuous in time sequence.

[0072] For the videos to be analyzed on video websites, since the audience of videos on video websites is in the hundreds of millions and the audience is spread across all age groups, according to the actual viewing behavior of users, corresponding in-video evaluation information (scores ranging from 0-100) and inter-video evaluation information (scores ranging from 0-+∞) will be generated for each second of playback in each video.

[0073] Each candidate video segment in the video to be analyzed has a corresponding first evaluation and a second evaluation.

[0074] The first evaluation of any candidate video segment is used to indicate the relationship between the viewing situation of the candidate video segment (for example, the number of plays, the number of comments, the number of bullet comments, etc.) and the viewing situation of other candidate video segments in the video to be analyzed. For example, the first evaluation of the candidate video segment I with the lowest first evaluation in the video to be analyzed is 0, and the first evaluation of the candidate video segment II with the highest first evaluation is 100; for any candidate video segment i, the corresponding first evaluation can be obtained based on the first viewing situation a with a first evaluation of 0 and the first viewing situation b with a first evaluation of 0 and a first evaluation of 100; exemplarily, when the parameter type included in the viewing situation only includes the number of plays, the first viewing situation a is played 1000 times, the first viewing situation b is played 11000 times, and the candidate video segment i is played 5000 times, then the first score of the candidate video segment i can be:

[0075] (100×5000) / (11000-1000)=50;

[0076] In addition, other methods and other parameter types of viewing conditions may be used to determine the first score of any candidate video segment, which are not listed here one by one.

[0077] The second evaluation of any candidate video segment is used to indicate the relationship between the viewing situation of the candidate video segment (for example, the number of plays, the number of comments, the number of comments, etc.) and the viewing situation of other video segments in the labeled video set (including other videos except the video to be analyzed). For example, the second evaluation of video segment a with the lowest second viewing situation a in the video to be analyzed is 0, and the second evaluation of video segment b with the second viewing situation b is 400; the second evaluation of any candidate video segment can be determined based on the viewing situation of any candidate video segment, the second evaluation of video segment a with the lowest second viewing situation a is 0, and the second evaluation of video segment b with the second viewing situation b is 400 with reference to the aforementioned method for determining the first rating.

[0078] Optionally, the intra-video evaluation information and the inter-video evaluation information may be obtained before the video to be processed is acquired. Therefore, the first target evaluation and the second target evaluation of the target video segment may be determined directly based on the intra-video evaluation information and the inter-video evaluation information. Step S103 : Determine a target excitement rating value corresponding to the target video segment based on the intra-video evaluation information, the inter-video evaluation information, the first target evaluation, and the second target evaluation.

[0079] After determining the intra-video evaluation information, inter-video evaluation information, first target evaluation, and second target evaluation, a target excitement evaluation value corresponding to the target video segment can be determined based on the intra-video evaluation information, inter-video evaluation information, first target evaluation, and second target evaluation. For example:

[0080] Determine the first exciting evaluation values ​​corresponding to the different first evaluations and the second exciting evaluation values ​​corresponding to the different second evaluations, determine the first exciting evaluation value corresponding to the first target evaluation and the second exciting evaluation value corresponding to the second target evaluation; and finally determine the target exciting evaluation value. Step S104: Determine the exciting category corresponding to the target video segment based on the target exciting evaluation value corresponding to the target video segment.

[0081] After the target wonderfulness evaluation value of the target video is determined, the wonderfulness category corresponding to the target video segment can be determined according to the target wonderfulness evaluation value.

[0082] For example, different exciting degree evaluation value intervals corresponding to different exciting degree categories may be preset, and then the exciting degree category corresponding to the target video segment may be determined according to the exciting degree evaluation value interval into which the target exciting degree evaluation value interval falls.

[0083] Furthermore, when determining the excitement category corresponding to the target video segment, the video content actually displayed in the target video segment may be determined, and the excitement category may be determined in combination with the video content and the target excitement evaluation value.

[0084] Through the method in this embodiment, an implementation method is provided for automatically determining the excitement category corresponding to a video clip. Compared with the manual labeling of the excitement of the video in the related art, it can effectively ensure the uniformity of the judgment standard and improve the efficiency of determining the excitement category, thereby effectively overcoming the technical problems of the manual labeling of the excitement of the video in the related art, which is highly subjective, slow, and unable to achieve fast batch data production.

[0085] As an optional embodiment, as in the aforementioned method, step S103 determines a target excitement evaluation value corresponding to the target video segment based on the intra-video evaluation information, the inter-video evaluation information, the first target evaluation, and the second target evaluation, including the following steps:

[0086] Step S201 : determining a first candidate evaluation corresponding to each candidate video segment according to the evaluation information within the video.

[0087] After obtaining the in-video evaluation information, which can generally be a curve of evaluation change, the in-video evaluation information can then be determined to determine a first candidate evaluation corresponding to each candidate video segment.

[0088] Step S202 : Calculate the average value of all first candidate evaluations to obtain a first average value.

[0089] After all first candidate evaluations are obtained, an average value of all first candidate evaluations may be calculated to obtain a first average value corresponding to the video to be analyzed.

[0090] Step S203, determining a second candidate evaluation corresponding to each candidate video segment according to the inter-video evaluation information;

[0091] After obtaining the intra-video evaluation information, the inter-video evaluation information can generally be a curve of evaluation changes (compared to the situation where the curve of the intra-video evaluation information has an upper limit (for example, 100), the inter-video evaluation information has no upper limit), and then, the second candidate evaluation corresponding to each candidate video segment can be determined.

[0092] Step S204: Calculate the average value of all second candidate evaluations to obtain a second average value.

[0093] After all second candidate evaluations are obtained, an average value of all second candidate evaluations may be calculated to obtain a second average value corresponding to the video to be analyzed.

[0094] Step S205: determining a first size relationship between the first target evaluation and the first average value; and determining a second size relationship between the second target evaluation and the second average value.

[0095] After obtaining the first target evaluation and the first average value, a first magnitude relationship between the first target evaluation and the first average value may be determined. As an optional embodiment, as in the aforementioned method, determining the first magnitude relationship between the first target evaluation and the first average value includes the following steps:

[0096] Step S301 : determining a first minimum candidate evaluation with the lowest evaluation value and a first maximum candidate evaluation with the highest evaluation value among all first candidate evaluations.

[0097] After all first candidate evaluations are determined, a first minimum candidate evaluation with the lowest evaluation value and a first maximum candidate evaluation with the highest evaluation value may be determined from among all first candidate evaluations by comparing them one by one.

[0098] Step S302 : dividing the first minimum candidate evaluation and the first average value to obtain a first preset number of first low evaluation value intervals; dividing the first highest candidate evaluation and the first average value to obtain a second preset number of first high evaluation value intervals.

[0099] After obtaining the first minimum candidate evaluation and the first average value, an evaluation interval with the first minimum candidate evaluation as the minimum value and the first average value as the maximum value can be obtained. The evaluation interval can be divided according to a first preset number of divisions to obtain a first preset number of first low evaluation value intervals. Optionally, the first preset number of first low evaluation value intervals can be obtained by uniform division. Furthermore, no two different first low evaluation value intervals have any intersection.

[0100] After obtaining the first highest candidate evaluation and the first average value, an evaluation interval with the first highest candidate evaluation as the maximum value and the first average value as the minimum value can be obtained. This evaluation interval can then be divided according to a second preset number of intervals to obtain a second preset number of first high evaluation value intervals. Optionally, a uniform division method can be used to obtain the second preset number of first high evaluation value intervals. Furthermore, no two different first high evaluation value intervals have any intersection.

[0101] Step S303, obtaining a first size relationship between the first target evaluation and the first average value by determining a first target evaluation value interval including the first target evaluation value in all first evaluation value intervals, wherein all first evaluation value intervals include a first low evaluation value interval and a first high evaluation value interval.

[0102] After obtaining the first low evaluation value interval and the first high evaluation value interval, all first evaluation value intervals can be obtained, and the first evaluation value interval in which the first target evaluation value falls can be determined as the first target evaluation value interval, and the first target evaluation value interval containing the first target evaluation value can be determined as the first size relationship between the first target evaluation and the first average value.

[0103] For example, when the first average value is 56, the first preset number is 7, and the second preset number is 4, the first low evaluation value intervals that can be obtained include [0, 8), [8, 16), [16, 24), [24, 32), [32, 40), [40, 48), and [48, 56); the first high evaluation value intervals include [56, 67), [67, 78), [78, 89), and [89, 100]. If the first target evaluation value is 66, the first high evaluation value interval [56, 67) is determined to be the first target evaluation value interval.

[0104] As an optional embodiment, as in the aforementioned method, determining the second magnitude relationship between the second target evaluation and the second average value includes the following steps:

[0105] Step S401 : determining the second minimum candidate evaluation with the lowest evaluation value and the second maximum candidate evaluation with the highest evaluation value among all second candidate evaluations.

[0106] After all second candidate evaluations are determined, the second minimum candidate evaluation with the lowest evaluation value and the second maximum candidate evaluation with the highest evaluation value can be determined from among all second candidate evaluations by comparing them one by one.

[0107] Step S402: divide the second minimum candidate evaluation and the second average value to obtain a third preset number of second low evaluation value intervals; determine the minimum specified second evaluation between the second highest candidate evaluation and the preset second evaluation upper limit, divide the specified second evaluation and the second average value to obtain a fourth preset number of second high evaluation value intervals.

[0108] After obtaining the second minimum candidate evaluation and the second average value, an evaluation interval with the second minimum candidate evaluation as the minimum value and the second average value as the maximum value can be obtained. The evaluation interval can be divided according to a third preset number of intervals to obtain a third preset number of second low evaluation value intervals. Optionally, the third preset number of second low evaluation value intervals can be obtained by uniform division. Furthermore, no two different second low evaluation value intervals have any intersection.

[0109] After obtaining the second highest candidate evaluation and the second average value, the evaluation value interval can be divided according to the second highest candidate evaluation and the second average value. Since the second highest candidate evaluation may have a peak value far higher than the average value, the minimum designated second evaluation can be determined between the second highest candidate evaluation and the preset second evaluation upper limit.

[0110] The preset second evaluation upper limit may be a pre-set evaluation upper limit, for example, 400, 500, etc.

[0111] For example, when the upper limit of the second evaluation is 400 and the second highest candidate evaluation is 300, 300 is used as the designated second evaluation; when the upper limit of the second evaluation is 400 and the second highest candidate evaluation is 500, 400 is used as the designated second evaluation.

[0112] After determining the designated second evaluation, an evaluation interval with the designated second evaluation as the maximum value and the second average value as the minimum value can be obtained. The evaluation interval can be divided according to a fourth predetermined number of intervals to obtain a fourth predetermined number of second-highest evaluation value intervals. Optionally, the fourth predetermined number of intervals can be divided evenly. Furthermore, no two different second-highest evaluation value intervals overlap.

[0113] Step S403, obtaining a second size relationship between the second target evaluation and the second average value by determining a second target evaluation value interval including the second target evaluation value in all second evaluation value intervals, wherein all second evaluation value intervals include a second low evaluation value interval and a second high evaluation value interval.

[0114] After obtaining the second lowest evaluation value interval and the second highest evaluation value interval, all second evaluation value intervals can be obtained, and the second evaluation value interval in which the second target evaluation value falls can be determined as the second target evaluation value interval, and the second target evaluation value interval containing the second target evaluation value can be determined as the first size relationship between the second target evaluation and the second average value.

[0115] For example, when the second average value is 70, the third preset number is 7, the second highest candidate evaluation is 310, the preset second evaluation upper limit is 400, and the fourth preset number is 4, the designated second evaluation can be determined to be 300, and the second lowest evaluation value intervals obtained include [0,10), [10,20), [20,30), [30,40), [40,50), [50,60), [60,70); the second highest evaluation value intervals include [70,130), [130,190), [190,250), [250,310]. If the second target evaluation is 66, the second lowest evaluation value interval [60,70) is determined as the second target evaluation value interval; if the second target evaluation is 199, the second highest evaluation value interval [190,250) is determined as the second target evaluation value interval.

[0116] Step S206 : determining a target excitement evaluation value corresponding to the target video segment based on the first size relationship and the second size relationship.

[0117] After determining the first size relationship and the second size relationship, a target excitement evaluation value corresponding to the target video segment can be determined according to the first size relationship and the second size relationship. As an optional embodiment, as in the aforementioned method, determining the target excitement evaluation value corresponding to the target video segment based on the first size relationship and the second size relationship includes the following steps:

[0118] Step S501 : determining a first wonderfulness value corresponding to each first evaluation value interval; and determining a second wonderfulness value corresponding to each second evaluation value interval.

[0119] After obtaining all first evaluation value intervals, a first excitement value can be assigned to each first evaluation value interval, thereby obtaining the first excitement value corresponding to each first evaluation value interval. For example, if there are 11 first evaluation value intervals, from low to high, the corresponding first excitement values ​​can be 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, and 10, respectively. Similarly, a second excitement value corresponding to each second evaluation value interval can be determined.

[0120] Step S502: Determine the first target wonderfulness value corresponding to the first target evaluation value interval according to the first wonderfulness value corresponding to each first evaluation value interval; determine the second target wonderfulness value corresponding to the second target evaluation value interval according to the second wonderfulness value corresponding to each second evaluation value interval.

[0121] After determining the first wonderfulness value corresponding to each first evaluation value interval, and the first target evaluation value interval is one of all the first evaluation value intervals, the first target wonderfulness value corresponding to the first target evaluation value interval can be determined.

[0122] Similarly, after determining the second wonderfulness value corresponding to each second evaluation value interval, and the second target evaluation value interval is one of all the second evaluation value intervals, the second target wonderfulness value corresponding to the second target evaluation value interval can be determined.

[0123] Step S503 : Calculating the first target excitement value and the second target excitement value according to a preset weighting method to obtain a target excitement evaluation value.

[0124] After obtaining the first target excitement value and the second target excitement value, the first target excitement value and the second target excitement value may be calculated according to a preset weighting method.

[0125] The preset weighting method can limit the first weight corresponding to the first target excitement value and the second weight corresponding to the second target excitement value. In general, the first weight and the second weight can both be 0.5, that is, the average value of the first target excitement value and the second target excitement value is calculated to obtain the target excitement evaluation value.

[0126] For example, when the first target excitement value is 7 and the second target excitement value is 9, and the preset weighting method is to calculate the average value, the obtained target excitement evaluation value is (7+9) / 2=8.

[0127] Step S504 : When the second target evaluation value is higher than or equal to the preset second evaluation upper limit, the target video segment is assigned a maximum target excitement evaluation value.

[0128] For clips whose second target evaluation is higher than the preset second evaluation upper limit (for example, 400, such video clips have an absolute number of positive viewing operations), the target video clip is assigned the maximum target wonderfulness evaluation value, for example, it is directly judged as a video clip with a wonderfulness of 10 points.

[0129] The method in this embodiment provides an implementation method for automatically calculating the target excitement evaluation value corresponding to the target video clip, which can effectively improve the efficiency of determining the target excitement evaluation value and avoid the technical problems of large subjective images and low efficiency caused by manual determination.

[0130] As an optional embodiment, as in the above method, step S105 determines the excitement category corresponding to the target video segment according to the target excitement evaluation value corresponding to the target video segment, including the following steps:

[0131] Step S601: Perform behavior detection on the target video clip to obtain a behavior detection result.

[0132] After the target video segment is obtained, behavior detection may be performed on the target video segment using a preset video behavior detection algorithm to obtain a behavior detection result corresponding to the target video segment.

[0133] The behavior detection result may be used to indicate the behavior indicated in the target video clip, such as skiing, fighting, laughing, crying, etc.

[0134] Step S602: If the behavior detection result indicates that a behavior of a preset behavior type exists in the target video segment and the target excitement evaluation value is greater than or equal to a preset threshold, determine that the excitement category corresponding to the target video segment is a high excitement category.

[0135] After the behavior detection result is determined, it can be determined whether the behavior detection result indicates that there is a behavior of a preset behavior type in the target video segment.

[0136] The preset behavior categories may be information indicating types of behaviors that may have actual exciting meanings, such as skiing, fighting, laughing, crying, etc. Moreover, the preset behavior categories may be added, deleted, or modified according to the behavior categories to be detected in actual applications.

[0137] After the target excitement evaluation value is determined, the relationship between the target excitement evaluation value and a preset threshold value may be determined.

[0138] The preset threshold can be a pre-set threshold for distinguishing between high excitement evaluation values ​​and low excitement evaluation values. For example, when the preset threshold is 6, if the target excitement evaluation value is greater than or equal to 6, the target excitement evaluation value is a high excitement evaluation value; otherwise, it is a low excitement evaluation value.

[0139] When the behavior detection result indicates that there is a preset behavior type in the target video clip, and the target exciting evaluation value is greater than or equal to the preset threshold, it means that it contains behaviors that have actual exciting significance, and the target exciting evaluation value is a high exciting evaluation value. This shows that the target video clip is not exciting because the video to be analyzed itself is exciting, but the target video clip itself does contain exciting behavior clips that make it exciting (for example, fighting scenes, quarreling scenes, kissing scenes, laughing, crying scenes, etc.). Therefore, the exciting category corresponding to the target video clip is determined to be the high exciting category.

[0140] Step S603 : if the behavior detection result indicates that the target video segment contains a behavior of a preset behavior type and the target excitement evaluation value is less than a preset threshold, determine that the excitement category corresponding to the target video segment is a low excitement category.

[0141] When the behavior detection result indicates that there is a preset behavior type in the target video clip, and the target excitement evaluation value is less than the preset threshold, it means that although the target video clip contains behaviors that are actually exciting, the target excitement evaluation value is still a low excitement evaluation value, which means that the excitement of the video to be detected is low. This is not because there is no behavior (for example, empty shot scenes, ordinary chat scenes, scenes where the actors have no language, no actions, no emotional expression, etc.), but because the behavior corresponding to the target video clip does not attract the audience, resulting in low excitement. Therefore, in this case, the excitement category corresponding to the target video clip is determined to be a low excitement category, which can be more in line with the user's viewing habits.

[0142] As an optional embodiment, as in the above method, after determining the excitement category corresponding to the target video segment according to the target excitement evaluation value corresponding to the target video segment in step S105, the method further includes the following steps:

[0143] Step S701 : determining a designated excitement category corresponding to each designated video segment, wherein all designated video segments include a target video segment.

[0144] Before step S701, a target request for generating training data for training the excitement algorithm can be obtained, and then all specified video clips can be obtained in response to the target request, and the specified excitement category corresponding to each specified video clip can be determined according to the method in the aforementioned embodiment.

[0145] Step S702: The designated video segment whose corresponding designated excitement category is the high excitement category is used as a high excitement segment.

[0146] Step S703: The designated video segment whose corresponding designated excitement category is the low excitement category is used as a low excitement segment.

[0147] Step S704 : selecting a first number of high-highlight training segments from all high-highlight segments and selecting a second number of low-highlight training segments from all low-highlight segments according to a preset quantity relationship, wherein the first number and the second number satisfy the preset quantity relationship.

[0148] After determining the designated excitement category corresponding to each designated video clip, all designated video clips can be classified according to the excitement category, and designated video clips whose corresponding designated excitement category is the high excitement category can be regarded as high excitement clips, and designated video clips whose corresponding designated excitement category is the low excitement category can be regarded as low excitement clips.

[0149] In order to facilitate the provision of training data for training the excitement algorithm in the later stage, a first number of high-excitement training segments can be selected from all high-excitement segments, and a second number of low-excitement training segments can be selected from all low-excitement segments according to a preset quantity relationship.

[0150] The preset quantity relationship may be used to indicate the proportional relationship between the number of high-excitement segments and the number of low-excitement segments in all training data, or directly limit the number of high-excitement segments and the number of low-excitement segments.

[0151] Furthermore, after determining the preset quantity relationship, a first quantity and a second quantity can be determined, and then the first quantity of high-highlight training segments are selected from all high-highlight segments, and the second quantity of low-highlight training segments are selected from all low-highlight segments.

[0152] The high-excitement training clips are positive sample data used to train the excitement algorithm, and the low-excitement training clips are negative sample data used to train the excitement algorithm.

[0153] In addition, training data for the excitement algorithm can be dynamically added based on the newly added video content to ensure that the training data can meet current needs.

[0154] By using the method in this embodiment, high-excitement training segments and low-excitement training segments for training an excitement algorithm can be quickly determined, thereby effectively improving the production efficiency of training data.

[0155] As described below, an application example of any of the above embodiments is provided:

[0156] (1) Convert the existing filter scores (i.e., intra-video evaluation information) and filter-related scores (inter-video evaluation information) into specific wonderfulness scores (i.e., wonderfulness evaluation values). The specific conversion algorithm is as follows:

[0157] First, both filter data and filter-related data have different score distributions and average scores for different videos. Filter-related data primarily describes the distribution of viewing times per second across the entire video, while filter-related data primarily describes the global distribution of viewing times per second (e.g., across all videos on the entire video platform). To analyze the distribution across all videos, both scores must be analyzed separately during the score conversion process.

[0158] like Figure 2 and Figure 3 The following table shows the filter scores and filter-related scores for two TV series (i.e., Video 1 and Video 2) on two video platforms. The horizontal axis represents each time point in seconds, and the vertical axis represents the score corresponding to that time point.

[0159] Through 2 and Figure 3 As can be seen, the score distribution varies greatly between videos, and there is no specific upper limit for the maximum score of filter-related scores (the score distribution is [0, +∞)). At the same time, for the same video, there will be certain differences in the distribution of its filter score and filter-related scores. Although there are differences, it can be determined that for each score distribution, positions with relatively high scores have more positive viewing operations (repeated playback, a large number of barrages, etc.). Positions with relatively low scores have more negative viewing operations (exiting the viewing, double-speed playback, a small number of barrages, etc.).

[0160] In view of the above distribution, a general score partitioning algorithm is proposed. The specific steps for processing the video to be analyzed are as follows:

[0161] 1. Calculate the first average score of the filter score in the video (i.e., the first average value) and the second average score of the filter-related score in the video (i.e., the second average value) avg_score, which are used to represent the average viewing situation of the video to be analyzed.

[0162] 2. For clips with filter-related data higher than 400 points (such clips have an absolute majority of positive viewing operations), they are directly judged as video clips with an excitement score of 10 points.

[0163] 3. By calculating the average score, all candidate video clips of the video to be analyzed are divided into two parts. Candidate video clips with higher scores than the average are considered to be highlights, while candidate video clips with lower scores than the average are considered to be ordinary clips. The calculation method is as follows:

[0164] First calculate the difference between the highest score and the average score delta_score 1 = max_score –

[0165] avg_score, delta_score1 is the score difference. If max_score is the maximum of the 5 values ​​of the filter data (i.e., the first highest candidate evaluation), its value is 100 points. If max_score is the filter

[0166] When the maximum value of the relevant data (i.e., specifying the second evaluation), max_score = min(400, max_relative_lvjing_score), where max_relative_lvjing_score is the second highest candidate evaluation in the aforementioned embodiment, and 400 is the second highest candidate evaluation in the aforementioned embodiment.

[0167] Then divide delta_score into 4 equal parts, and the value of each part is one_part0 (i.e., the first highest evaluation value interval in the filter data, or the second highest evaluation value interval in the filter related data).

[0168] Evaluation value interval) is one_part = delta_score 1 / 4. For [avg_score + 0 ×

[0169] The clips with scores in the range of [one_part,avg_score+1×one_part) are scored as wonderful and 7 points are given.

[0170] The clips in the range of [avg_score+2×one_part,avg_score5+3×one_part) are scored as wonderful with 8 points, the clips in the range of [avg_score+2×one_part,avg_score5+3×one_part) are scored as wonderful with 9 points, and the clips in the range of [avg_score

[0171] +3×one_part,max_score] is recorded as a wonderful score of 10 points. Because both the filter score and the filter-related score can be calculated to obtain the corresponding wonderful score, the clips with a wonderful score of 7 or above are regarded as the final wonderful clips, and their scores are calculated based on the value of the filter score.

[0172] Use the average of the two scores (ie, the default weighting method is the average calculation method): 0final_score=(score1+score2) / 2; score1 is the first target excitement value, score2

[0173] is the second target excitement value.

[0174] For video clips with scores below the average, first calculate the difference between the average score and the minimum score delta_score2 = avg_score – min_score. Here, min_score is the filter

[0175] The lowest score of the video in the filter data or filter-related data (i.e., the first 5 smallest candidate evaluations in the filter data, or the second smallest candidate evaluations in the filter-related data). Then delta_score2 is divided into 8 equal parts, and the value of each part is one_part=

[0176] delta_score2 / 8, for scores between [min_score+0×one_part, min_score+1×one_part), the excitement is recorded as 0 points. Similarly, the scores corresponding to videos with excitement of 1-5 points are obtained using the formula [min_score+i×one_part, min_score+(i+1)×one_part), where i represents the specific excitement score. The score corresponding to a video with an excitement of 6 points is obtained using [min_score+6×one_part, avg_score]. Similarly, because both the filter score and the filter-related score can be calculated to obtain the corresponding excitement score, for clips with an excitement of 6 points or less as the final non-exciting clips, their score uses the average of the two scores (that is, the preset weighting method is the average calculation method): final_score=(score1+score2) / 2.

[0177] 4. Determine whether the video clips with high scores contain actions that are actually exciting and meaningful.

[0178] For the high-scoring videos obtained from the video to be analyzed, video clips with a wonderful score higher than 7 points are identified using an existing video behavior detection algorithm (i.e., an algorithm for behavior detection), and information such as the behavior label (i.e., behavior detection result) and confidence level (which can be set according to the actual usage scenario, such as 95%, 90%, etc.) in each video clip is obtained. By using a higher confidence threshold and the behavior labels that need to be retained, video clips that have no actual wonderful meaning are filtered out.

[0179] Moreover, when there are new requirements for the excitement analysis algorithm, the corresponding training sample data can be produced dynamically and quickly. For example, when it is necessary to add the behavior type of "skiing" as the output content of the excitement algorithm, the video behavior detection algorithm can be used to quickly produce potential data with the behavior type of "skiing" for training the excitement algorithm.

[0180] For video clips with low excitement scores (i.e., the target excitement evaluation value is less than the preset threshold), if they still contain behavioral clips with actual exciting meaning, they are still considered to be low-excitement video clips even if they contain label data obtained by the above-mentioned video behavior detection algorithm. This is more in line with the user's viewing habits.

[0181] When training a highlight algorithm, the resulting data, when trained on large amounts of video data, is uneven across different highlight scores. This is often because, after filtering using the behavior recognition algorithm, high-score segments are significantly lower than low-score segments. Here, we simply retain as many high-score highlight segments as possible, while also retaining some of the low-score segments, maintaining a roughly 1:1 ratio (i.e., a pre-set relationship) for each highlight score.

[0182] like Figure 4 As shown, according to an embodiment of another aspect of the present application, a video analysis device is further provided, including:

[0183] Acquisition module 1, used to acquire the video to be analyzed;

[0184] A first determination module 2 is configured to determine a first target evaluation and a second target evaluation corresponding to a target video segment according to intra-video evaluation information and inter-video evaluation information corresponding to the video to be analyzed, wherein the intra-video evaluation information includes the first target evaluation corresponding to the target video segment, the inter-video evaluation information includes the second target evaluation corresponding to the target video segment, the first target evaluation is determined according to the evaluation of each candidate video segment in the video to be analyzed, and the second target evaluation is a global evaluation determined according to the evaluation of each video segment in a target video set, all candidate video segments include the target video segment, each candidate video segment has a unique corresponding time period in the video to be analyzed, and the target video set includes the video to be analyzed;

[0185] A second determining module 3 is configured to determine a target wonderfulness evaluation value corresponding to a target video segment based on the intra-video evaluation information, the inter-video evaluation information, the first target evaluation, and the second target evaluation;

[0186] The wonderfulness category determination module 4 is configured to determine the wonderfulness category corresponding to the target video segment according to the target wonderfulness evaluation value corresponding to the target video segment.

[0187] Specifically, the specific process of each module in the device of the embodiment of the present invention realizing its function can be referred to the relevant description in the method embodiment, which will not be repeated here.

[0188] According to another embodiment of the present application, there is also provided an electronic device, including: Figure 5 As shown, the electronic device may include: a processor 1501 , a communication interface 1502 , a memory 1503 and a communication bus 1504 , wherein the processor 1501 , the communication interface 1502 , and the memory 1503 communicate with each other via the communication bus 1504 .

[0189] Memory 1503, used for storing computer programs;

[0190] The processor 1501 is configured to implement the steps of the above method embodiment when executing the program stored in the memory 1503 .

[0191] The bus mentioned in the electronic device mentioned above may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. This bus can be divided into address buses, data buses, control buses, etc. For ease of illustration, only one thick line is used in the figure, but this does not mean that there is only one bus or only one type of bus.

[0192] The communication interface is used for communication between the above electronic device and other devices.

[0193] The memory may include random access memory (RAM) or non-volatile memory (NVM), such as at least one disk storage. Alternatively, the memory may be at least one storage device located away from the processor.

[0194] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, and discrete hardware components.

[0195] An embodiment of the present application further provides a computer-readable storage medium, the storage medium including a stored program, wherein the method steps of the above method embodiment are executed when the program is run.

[0196] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or device comprising the element.

[0197] The foregoing description is intended only to provide specific embodiments of the present invention, which will enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is intended to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A video analysis method, characterized in that: include: Get the video to be analyzed; According to the in-video evaluation information and inter-video evaluation information corresponding to the video to be analyzed, a first target evaluation and a second target evaluation corresponding to the target video segment are determined, wherein the in-video evaluation information includes the first target evaluation corresponding to the target video segment, and the inter-video evaluation information includes the second target evaluation corresponding to the target video segment. The first target evaluation is determined according to the evaluation of each candidate video segment in the video to be analyzed, and is used to indicate the relationship between the viewing situation of the target video segment and the viewing situation of other candidate video segments in the video to be analyzed. The second target evaluation is a global evaluation determined according to the evaluation of each video segment in the target video set, and is used to indicate the relationship between the viewing situation of the target video segment and the viewing situation of other video segments in the target video set. The target video segment is included in all candidate video segments, and each candidate video segment has a unique corresponding time period in the video to be analyzed. The target video set includes the video to be analyzed. Based on the intra-video evaluation information, the inter-video evaluation information, the first target evaluation and the second target evaluation, a target excitement evaluation value corresponding to the target video segment is determined, including: determining a first candidate evaluation corresponding to each candidate video segment according to the intra-video evaluation information; calculating the average value of all first candidate evaluations to obtain a first average value; determining a second candidate evaluation corresponding to each candidate video segment according to the inter-video evaluation information; calculating the average value of all second candidate evaluations to obtain a second average value; determining a first size relationship between the first target evaluation and the first average value; determining a second size relationship between the second target evaluation and the second average value; and determining a target excitement evaluation value corresponding to the target video segment based on the first size relationship and the second size relationship. According to the target excitement evaluation value corresponding to the target video segment, an excitement category corresponding to the target video segment is determined.

2. The method according to claim 1, characterized in that Determining a first magnitude relationship between the first target evaluation and the first average value includes: Determining a first minimum candidate evaluation with the lowest evaluation value and a first maximum candidate evaluation with the highest evaluation value among all first candidate evaluations; Dividing the first minimum candidate evaluation and the first average value to obtain a first preset number of first low evaluation value intervals; dividing the first maximum candidate evaluation and the first average value to obtain a second preset number of first high evaluation value intervals; By determining a first target evaluation value interval including the first target evaluation value in all first evaluation value intervals, a first size relationship between the first target evaluation and the first average value is obtained, wherein all first evaluation value intervals include the first low evaluation value interval and the first high evaluation value interval.

3. The method according to claim 2, characterized in that Determining a second magnitude relationship between the second target evaluation and the second average value includes: Determining a second minimum candidate evaluation with the lowest evaluation value and a second maximum candidate evaluation with the highest evaluation value among all second candidate evaluations; Dividing the second minimum candidate evaluation and the second average value to obtain a third preset number of second low evaluation value intervals; determining the minimum designated second evaluation between the second highest candidate evaluation and the preset second evaluation upper limit, and dividing the designated second evaluation and the second average value to obtain a fourth preset number of second high evaluation value intervals; A second size relationship between the second target evaluation and the second average value is obtained by determining a second target evaluation value interval including the second target evaluation value in all second evaluation value intervals, wherein all second evaluation value intervals include the second low evaluation value interval and the second high evaluation value interval.

4. The method according to claim 3, characterized in that The determining of a target excitement evaluation value corresponding to the target video segment based on the first size relationship and the second size relationship includes: Determine a first wonderfulness value corresponding to each first evaluation value interval; determine a second wonderfulness value corresponding to each second evaluation value interval; Determining a first target exciting value corresponding to the first target exciting value interval according to the first exciting value corresponding to each first evaluation value interval; determining a second target exciting value corresponding to the second target exciting value interval according to the second exciting value corresponding to each second evaluation value interval; Obtaining the target excitement evaluation value by calculating the first target excitement value and the second target excitement value according to a preset weighting method; When the second target evaluation value is higher than or equal to the preset second evaluation upper limit, the target video segment is assigned the maximum target excitement evaluation value.

5. The method according to claim 1, wherein The determining of the excitement category corresponding to the target video segment according to the target excitement evaluation value corresponding to the target video segment includes: Performing behavior detection on the target video clip to obtain a behavior detection result; If the behavior detection result indicates that a behavior of a preset behavior type exists in the target video segment, and the target excitement evaluation value is greater than or equal to a preset threshold, determining that the excitement category corresponding to the target video segment is a high excitement category; When the behavior detection result indicates that a behavior of a preset behavior type exists in the target video segment and the target excitement evaluation value is less than a preset threshold, the excitement category corresponding to the target video segment is determined to be a low excitement category.

6. The method according to any one of claims 1 to 5, characterized in that After determining the excitement category corresponding to the target video segment according to the target excitement evaluation value corresponding to the target video segment, the method further includes: Determining a designated excitement category corresponding to each designated video segment, wherein all designated video segments include the target video segment; The designated video segment corresponding to the designated excitement category being a high excitement category is used as a high excitement segment; The designated video segment corresponding to the designated excitement category being a low excitement category is used as a low excitement segment; According to a preset quantity relationship, a first number of high-highlight training segments are selected from all the high-highlight segments, and a second number of low-highlight training segments are selected from all the low-highlight segments, wherein the first number and the second number satisfy the preset quantity relationship.

7. A video analysis device, characterized in that: include: An acquisition module is used to acquire the video to be analyzed; A first determination module is configured to determine a first target evaluation and a second target evaluation corresponding to a target video segment according to intra-video evaluation information and inter-video evaluation information corresponding to the video to be analyzed, wherein the intra-video evaluation information includes the first target evaluation corresponding to the target video segment, and the inter-video evaluation information includes the second target evaluation corresponding to the target video segment. The first target evaluation is determined according to the evaluation of each candidate video segment in the video to be analyzed, and is used to indicate the relationship between the viewing situation of the target video segment and the viewing situation of other candidate video segments in the video to be analyzed. The second target evaluation is a global evaluation determined according to the evaluation of each video segment in a target video set, and is used to indicate the relationship between the viewing situation of the target video segment and the viewing situation of other video segments in the target video set. The target video segment is included in all candidate video segments, each candidate video segment has a unique corresponding time period in the video to be analyzed, and the target video set includes the video to be analyzed. The second determination module is used to determine the target excitement evaluation value corresponding to the target video segment based on the intra-video evaluation information, the inter-video evaluation information, the first target evaluation and the second target evaluation, including: determining the first candidate evaluation corresponding to each candidate video segment according to the intra-video evaluation information; calculating the average value of all the first candidate evaluations to obtain a first average value; determining the second candidate evaluation corresponding to each candidate video segment according to the inter-video evaluation information; calculating the average value of all the second candidate evaluations to obtain a second average value; determining a first size relationship between the first target evaluation and the first average value; determining a second size relationship between the second target evaluation and the second average value; and determining the target excitement evaluation value corresponding to the target video segment based on the first size relationship and the second size relationship. The excitement category determination module is configured to determine an excitement category corresponding to the target video segment according to a target excitement evaluation value corresponding to the target video segment.

8. An electronic device, characterized in that: include: A processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus; The memory is used to store computer programs; The processor is configured to implement the method according to any one of claims 1 to 6 when executing the computer program.

9. A computer-readable storage medium, characterized in that The storage medium includes a stored program, wherein the program executes the method according to any one of claims 1 to 6 when executed.

Images