Video pushing method and device, storage medium and electronic device

Abstract

The application discloses a video pushing method and device, a storage medium and electronic equipment. The method comprises the following steps: acquiring facial expression information of a target object watching a history video, attention information of the target object to the history video and interaction information of the target object and the history video in the process of playing the history video; determining a target video according to the facial expression information, the attention information and the interaction information; and pushing the target video to the target object. The application solves the technical problem of poor course pushing accuracy.

Description

Technical Field

[0001] This invention relates to the field of video push, and more specifically, to a video push method, apparatus, storage medium, and electronic device. Background Technology

[0002] In the early stages of children's programming education, programming courses should primarily be guided by children's interests. All course design should begin with understanding children. Only in this way can courses truly stimulate children's interest, cultivate their creativity and desire to explore, help them build positive feelings, and form a virtuous cycle in which children love to learn and gain knowledge. Truly understanding children's interests is the key to solving this problem.

[0003] In existing technologies, course recommendations rely entirely on human experience and cannot effectively recommend courses that are of interest to children and beneficial to their learning. Summary of the Invention

[0004] This invention provides a video push method, apparatus, storage medium, and electronic device to at least solve the technical problem of poor accuracy in pushing courses.

[0005] According to one aspect of the present invention, a video push method is provided, comprising: acquiring facial expression information of a target object watching the historical video, attention information of the target object to the historical video, and interaction information between the target object and the historical video during the playback of historical videos; determining a target video based on the facial expression information, the attention information, and the interaction information; and pushing the target video to the target object.

[0006] According to another aspect of the present invention, a video push device is provided, comprising: an acquisition module, configured to acquire facial expression information of a target object watching the historical video, attention information of the target object to the historical video, and interaction information between the target object and the historical video during the playback of a historical video; a determination module, configured to determine a target video based on the facial expression information, the attention information, and the interaction information; and a push module, configured to push the target video to the target object.

[0007] As an optional example, the determination module includes: a first determination unit, configured to determine the target object's interest score in the historical video based on the facial expression information, the attention information, and the interaction information; and a second determination unit, configured to determine the target video based on the interest score.

[0008] As an optional example, the first determining unit includes: a determining subunit, used to determine a first score of the facial expression information, a second score of the attention information, and a third score of the interaction information; and to determine the weighted sum of the first score, the second score, and the third score as the interest score.

[0009] As an optional example, the facial expression information includes facial images of the target object during the process of watching the historical video. The determining subunit is further configured to: input the facial images into the expression recognition model, and have the expression recognition model identify the emotion type of the target object; count the number of occurrences of each emotion type of the target object; and determine the first score as the ratio of the number of occurrences of positive emotions to the total number of emotions among the emotion types.

[0010] As an optional example, the aforementioned attention information includes the number of screen-switching operations performed by the target object, the number of skip operations performed by the target object, the number of unattended screens of the target object, and the total duration of the historical videos. The determining subunit is further configured to: substitute the number of screen-switching operations, the number of skip operations, the number of unattended screens, and the total duration into the first scoring formula to obtain the second score, wherein the first scoring formula is used to calculate the second score, and the second score is negatively correlated with the total duration.

[0011] As an optional example, the interactive information includes the correctness data, completion time, and number of submissions of the target object to the related questions of the historical video. The determining subunit is further used to: substitute the correctness data, completion time, and number of submissions into the second score formula to obtain the second score, wherein the second score formula is used to calculate the third score, and the correctness data is positively correlated with the third score.

[0012] As an optional example, the second determining unit includes: a selection subunit, configured to determine a first video as the target video, wherein the first video is a video of the same type as a historical video with an interest score higher than a first threshold; or to determine a second video as the target video, wherein the second video is a video of a similar type to a historical video with an interest score higher than the first threshold; or to determine a third video as the target video, wherein the third video is a video of the same type as the historical video with the highest interest score; or to determine a fourth video as the target video, wherein the fourth video is a video of a similar type to the historical video with the highest interest score.

[0013] As an optional example, the above-mentioned push module includes: a push unit, configured to push the target video to the target object so that the target object starts playing the target video after the current video has finished playing; or to push the target video to the target object and place the playback order of the target video after the video currently being played by the target object.

[0014] As an optional example, the second determining unit includes: a grouping subunit, used to determine the interest score of each target object to the historical video when there are multiple target objects, to obtain an interest score matrix; to group target objects with similar interest scores to the same historical video in the interest score matrix into an object group; and to determine the videos that other objects in the object group have watched and have high interest scores as the target videos.

[0015] According to another aspect of the present invention, a storage medium is also provided, wherein a computer program is stored in the storage medium, and the computer program is executed by a processor to perform the above-described video push method.

[0016] According to another aspect of the present invention, an electronic device is also provided, including a memory and a processor, wherein the memory stores a computer program and the processor is configured to execute the video push method described above through the computer program.

[0017] In this embodiment of the invention, a method is employed to acquire facial expression information of a target object watching a historical video, the target object's attention to the historical video, and the target object's interaction information with the historical video during playback; determine a target video based on the facial expression information, the attention information, and the interaction information; and push the target video to the target object. Because the target video is determined based on the target object's facial expression information, attention information, and interaction information during historical video playback, the accuracy of the target video is relatively high, conforming to the target object's viewing experience. This achieves the goal of improving the accuracy of video push, thereby solving the technical problem of poor accuracy in course push. Attached Figure Description

[0018] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0019] Figure 1This is a flowchart of an optional video push method according to an embodiment of the present invention;

[0020] Figure 2 This is a video playback diagram of an optional video push method according to an embodiment of the present invention;

[0021] Figure 3 This is a schematic diagram of the structure of an optional video push device according to an embodiment of the present invention;

[0022] Figure 4 This is a schematic diagram of an optional electronic device according to an embodiment of the present invention. Detailed Implementation

[0023] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0024] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0025] According to a first aspect of the present invention, a video push method is provided, optionally, as follows: Figure 1 As shown, the above method includes:

[0026] S102, during the playback of historical videos, obtain the facial expression information of the target object watching the historical videos, the target object's attention to the historical videos, and the interaction information between the target object and the historical videos.

[0027] S104, Determine the target video based on facial expression information, attention information, and interaction information;

[0028] S106, push the target video to the target object.

[0029] The aforementioned historical videos and target videos can be multimedia resources pushed to target objects or other objects. Videos can include video resources and associated resources. Video resources are playable resources, while associated resources can be interactive or manipulable resources, such as homework exercises or asking questions by raising hands.

[0030] When playing historical videos, the system can obtain facial expression information, attention information, and interaction information between the target audience and the video. Based on these three factors, it can determine the next target video to push and then push the target video to the target audience.

[0031] For example, taking the push of instructional videos to users as an example, before pushing, instructional video segments are prepared. Through surveys of user interests and educational research, instructional video segments are compiled based on different interest points, such as music, sports, games, drawing, and science. Corresponding practice exercises are also prepared so that suitable content can be retrieved from the instructional library promptly after the system identifies the user's interests. The video lengths can be the same or different. The number of practice exercises can be specified or not required. As an example, videos of the same type can be set to have the same duration, and the number of practice exercises can be set to the same.

[0032] After preparing the video clips, you can randomly select a video clip to play. Then, based on the user's reaction, collect the aforementioned facial expression information, the target's attention information, and the target's interaction information with the video to determine the user's level of liking for the video. Finally, based on the user's level of liking for various types of videos, decide on the next target video to be pushed.

[0033] In the above method, when pushing a target video to the target audience, the target video is determined based on the target audience's facial expression information, attention information of the target audience to the historical videos, and interaction information between the target audience and the historical videos during the playback of historical videos. Therefore, the accuracy of the target video is relatively high, which is in line with the target audience's viewing experience. This achieves the goal of improving the accuracy of video push and solves the technical problem of poor accuracy in pushing courses.

[0034] In this embodiment, when determining the target video, the target audience's interest score in historical videos can be determined based on facial expression information, attention information, and interaction information; the target video is then determined based on the interest score.

[0035] In other words, facial expression information, attention information, and interaction information are converted into quantifiable interest scores, and the level of the interest scores determines whether a video should be pushed as a target video.

[0036] When determining the interest score, a first score can be determined based on facial expression information, a second score based on attention information, and a third score based on interaction information. The weighted sum of the first, second, and third scores is then used to determine the interest score. The weights in the weighted summation process can be determined based on historical experience.

[0037] When calculating the first score, facial expression information of the user can be collected. The facial expression information can be a facial image. The facial image is input into the expression recognition model, which identifies the emotion type of the target object. The number of occurrences of each emotion type of the target object is counted. The ratio of the number of occurrences of positive emotions to the total number of emotions is determined as the first score.

[0038] In this embodiment, calculating the first fraction can be divided into the following steps:

[0039] 1. Use a camera to capture the user's facial image;

[0040] 2. Perform expression classification and analysis on the collected facial images;

[0041] 3. Select the expression value with the highest probability after expression classification and analysis as the user's facial expression;

[0042] 4. Collect facial expression data related to the course content, including: user ID, teaching video category, facial expression category and frequency, content viewer ID, and viewing time, for data analysis.

[0043] The facial images captured in step 1 above can be collected at predetermined intervals, or when the user's facial expressions change significantly. For example, if the user's facial expressions did not change significantly 3 minutes ago, but changed significantly at 3 minutes, exceeding a predetermined range, then a facial image of the user can be captured at 3 minutes. Furthermore, the user's facial expressions can be recorded in real time while watching each different type of segment.

[0044] The facial expression classification algorithm in step 2 above is based on a CNN deep learning framework, which trains a model to predict input facial expressions. It employs a multi-classification method, classifying an input facial expression image into six different emotional categories (happy, angry, fear, sad, surprised, and neutral). The emotion type is defined as `Emotion = max(P(img,e))`, where `img` is the face image, `e` represents the six different emotion categories (happy, angry, fear, sad, surprised, and neutral), and `P(img,e)` is the probability of classifying the face image into one of these six categories, an array containing six probability values ​​from 0 to 1. The `Emotion` returned to the client is the maximum value from this probability array. The specific CNN model construction process is as follows:

[0045] ① The network consists of 3 convolutional layers (C1, C2, C3), 3 maxpooling layers (S1, S2, S3), 1 flatten layer, 2 fully connected layers (FC1, FC2), 2 dropout layers, and 1 softmax layer.

[0046] ②The input layer is a 48*48 face pixel matrix. Before convolution, a ZeroPadding layer is applied to fill the boundaries of the image with 0s to ensure that the image size remains unchanged after convolution.

[0047] ③ The convolutional and pooling layers have several feature maps, each of which is connected to the feature map of the previous layer through local connections. Convolutional layers C1, C2, and C3 use 32, 64, and 128 convolutions respectively, and the kernel size of each convolutional layer is 3×3.

[0048] ④ The sampling window size used by pooling layers S1 and s2 is 2×2.

[0049] ⑤ The fully connected layer used to have a Flatten layer, which flattens the rolled-out three-dimensional layer into a one-dimensional layer to facilitate full connectivity.

[0050] ⑥ During training, a certain percentage of input neurons are randomly disconnected from the Dropout layer each time parameters are updated to prevent overfitting.

[0051] ⑦ The Softmax layer contains 6 neurons and classifies the features output by the fully connected layer, dividing facial expressions into 6 categories: happy, surprised, angry, sad, fearful, and neutral.

[0052] Specifically, the training set consists of 35,887 cropped, 48×48 pixel grayscale images of human faces, each labeled with one of six emotion classes: anger, fear, happiness, sadness, surprise, and neutral. After training the CNN model, a classification model file is obtained.

[0053] The facial expression data collected in step 4 above, namely the number of times a user is happy (x1), surprised (x2), angry (x3), sad (x4), fearful (x5), and neutral (x6) when watching a certain teaching segment, can provide a relatively clear understanding of user preferences for programming education. Starting from interests, understanding user needs, and judging the user's level of liking for the teaching content (x) are shown in the following formula:

[0054] X = (x1 + x2 + x6) / (x1 + x2 + x3 + x4 + x5 + x6), which means that the user's liking for the teaching segment can be determined based on the user's facial emotion recognition, and it is a value between 0 and 1.

[0055] When calculating the second score, the number of screen-switching operations performed by the target object, the number of skip operations performed by the target object, the number of unfollowed screens of the target object, and the total duration of historical videos can be obtained as attention information to determine the second score. Substituting the number of screen-switching operations, skip operations, unfollowed screens, and total duration into the first score formula yields the second score. The first score formula is used to calculate the second score, and the second score is negatively correlated with the total duration. The main steps are as follows:

[0056] 1. Record the user's attention performance for each segment, with specific dimensions including: number of screen switches y1, number of skips y2, number of times not on screen y3, and total video duration y4.

[0057] 2. Based on these dimensions, an analysis of user attention to the teaching content was obtained, showing a negative correlation between screen switching frequency, skipping frequency, off-screen frequency, total video duration, and attention level.

[0058] 3. By assigning weights to the three factors based on historical experience, we can obtain the user's level of focus on the teaching content, a value between 0 and 1. A higher value indicates higher focus, and a lower value indicates lower focus. The first formula is as follows:

[0059] Y = 1 - e^(-a*y1 - b*y2 - c*y3 - d*y4)

[0060] User-initiated behaviors, such as the number of screen switches and skips, better reflect a user's level of disinterest in the video. Therefore, the coefficients in the above formula are set as a=0.3, b=0.3, c=0.2, and d=0.2. The resulting Y is the second score.

[0061] The calculation of the third score involves obtaining the target audience's correct / incorrect answers to questions related to historical videos, completion time, and number of submissions as interaction information. These data are then substituted into the second score formula to obtain the second score, which is used to calculate the third score. The correct / incorrect answers are positively correlated with the third score. The main steps are as follows:

[0062] 1. Record the user's level completion status for each teaching segment, with specific dimensions including: level correctness z1, level completion time z2, and number of submissions z3.

[0063] 2. The user's acceptance of the teaching content is positively correlated with the correctness of the level, and negatively correlated with the level completion time and the number of submissions.

[0064] 3. By assigning weights to the three factors based on historical experience, we can obtain the user's acceptability of the teaching content. This value is between 0 and 1, with a larger value indicating higher user acceptability and a smaller value indicating lower acceptability.

[0065] In summary, the user's interest score for a video clip can be composed of three aspects: [facial emotion X, focus Y, and acceptability of the level practice Z]. Based on this, we can fit a specific formula for the user's interest in the teaching clip, namely: score = (X, Y, Z) = k1*X + k2*Y + k3*Z = 0.55*[(x1+x2+x6) / (x1+x2+x3+x4+x5+x6)] + 0.225*[1-e^(-a*y1-b*y2-c*y3-d*y4)] + 0.225*[e^(z1+z2-z3) / (1 + e^(z1+z2-z3))]. Since facial expressions contain the most information (up to 55%), K1 = 0.55, k2 = k3 = 0.225. It should be noted that the above parameters are not unique and can be adjusted. Simply ensure that the coefficients for facial expression parameters are higher than those for focus and interaction parameters.

[0066] Based on the user's interest score for each teaching segment, we can obtain the following user interest score matrix for the teaching segments:

[0067]

[0068] In this embodiment, since the score of the target object for each historical video can be determined, when determining the target video, videos of the same or similar type as the historical video with a score exceeding a first threshold or the highest score can be pushed to the target object as target videos. Therefore, the pushed target videos are videos of the same or similar type as the videos the target object likes.

[0069] After the target video is pushed to the target audience, the target audience can either watch the target video after finishing the current video, or end the current video playback and start playing the target video as soon as it is pushed. In this case, "once pushed" means the target audience has either fully downloaded the target video or has buffered a portion of it.

[0070] In this embodiment, an interest score matrix as described above can be determined for each user. Target objects with similar interest scores for the same historical videos in the interest score matrix are grouped into an object group; videos viewed by other users in the object group and with high interest scores are identified as target videos.

[0071] For example, if user A and user B have similar interest scores for the same historical videos, then for a video that user A has not watched but user B has watched and that user B likes, that video can be pushed to user A.

[0072] If a user's interest scores for all videos are low, a notification message will be sent to the associated object of that video, indicating that the user may be having difficulty learning and needs to pay special attention.

[0073] Figure 2 This is an exemplary diagram illustrating a user watching a target video. Figure 2 In the video playback area 202, the video list is displayed in the list area 204, and area 206 is the interactive area. Figure 2 The viewing interface shown allows users to fast forward, rewind, go to the next page, go back, and interact with the audience through quizzes. It should be noted that... Figure 2 This is for illustrative purposes only and does not show all the content, nor does it constitute a limitation.

[0074] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the present invention is not limited to the described order of actions, because according to the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to the present invention.

[0075] According to another aspect of the embodiments of this application, a video push device is also provided, such as... Figure 3 As shown, it includes:

[0076] The acquisition module 302 is used to acquire facial expression information of the target object watching the historical video, the target object's attention to the historical video, and the interaction information between the target object and the historical video during the playback of the historical video.

[0077] The determination module 304 is used to determine the target video based on facial expression information, attention information, and interaction information.

[0078] The push module 306 is used to push target videos to the target object.

[0079] The aforementioned historical videos and target videos can be multimedia resources pushed to target objects or other objects. Videos can include video resources and associated resources. Video resources are playable resources, while associated resources can be interactive or manipulable resources, such as homework exercises or asking questions by raising hands.

[0080] When playing historical videos, the system can obtain facial expression information, attention information, and interaction information between the target audience and the video. Based on these three factors, it can determine the next target video to push and then push the target video to the target audience.

[0081] For example, taking the push of instructional videos to users as an example, before pushing, instructional video segments are prepared. Through surveys of user interests and educational research, instructional video segments are compiled based on different interest points, such as music, sports, games, drawing, and science. Corresponding practice exercises are also prepared so that suitable content can be retrieved from the instructional library promptly after the system identifies the user's interests. The video lengths can be the same or different. The number of practice exercises can be specified or not required. As an example, videos of the same type can be set to have the same duration, and the number of practice exercises can be set to the same.

[0082] After preparing the video clips, you can randomly select a video clip to play. Then, based on the user's reaction, collect the aforementioned facial expression information, the target's attention information, and the target's interaction information with the video to determine the user's level of liking for the video. Finally, based on the user's level of liking for various types of videos, decide on the next target video to be pushed.

[0083] In the above method, when pushing a target video to the target audience, the target video is determined based on the target audience's facial expression information, attention information of the target audience to the historical videos, and interaction information between the target audience and the historical videos during the playback of historical videos. Therefore, the accuracy of the target video is relatively high, which is in line with the target audience's viewing experience. This achieves the goal of improving the accuracy of video push and solves the technical problem of poor accuracy in pushing courses.

[0084] For other examples of this embodiment, please refer to the examples above, which will not be repeated here.

[0085] Figure 4This is a structural block diagram of an optional electronic device according to an embodiment of this application, such as... Figure 4 As shown, it includes a processor 402, a communication interface 404, a memory 406, and a communication bus 408. The processor 402, communication interface 404, and memory 406 communicate with each other via the communication bus 408.

[0086] Memory 406 is used to store computer programs;

[0087] When processor 402 executes a computer program stored in memory 406, it performs the following steps:

[0088] During the playback of historical videos, information such as the facial expressions of the target audience watching the historical videos, the target audience's attention to the historical videos, and the interaction information between the target audience and the historical videos can be obtained.

[0089] Target videos are determined based on facial expression information, attention information, and interaction information.

[0090] Push the target video to the target audience.

[0091] Optionally, in this embodiment, the communication bus can be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. This communication bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 4 The symbol is represented by a single thick line, but this does not indicate that there is only one bus or one type of bus. The communication interface is used for communication between the aforementioned electronic devices and other devices.

[0092] The memory may include RAM, or non-volatile memory, such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0093] As an example, the memory 406 described above may include, but is not limited to, the first receiving unit 902, the first query unit 904, and the first sending unit 906 in the video push device described above. Furthermore, it may include, but is not limited to, other module units in the request processing device described above, which will not be elaborated upon in this example.

[0094] The processor mentioned above can be a general-purpose processor, including but not limited to: CPU (Central Processing Unit), NP (Network Processor), etc.; it can also be DSP (Digital Signal Processor), ASIC (Application Specific Integrated Circuit), FPGA (Field-Programmable Gate Array) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0095] Optionally, specific examples in this embodiment can refer to the examples described in the above embodiments, and will not be repeated here.

[0096] Those skilled in the art will understand that Figure 4 The structure shown is for illustrative purposes only. The device implementing the above video push method can be a terminal device, such as a smartphone (e.g., Android phone, iOS phone), tablet computer, PDA, mobile internet device (MID), PAD, etc. Figure 4 This does not limit the structure of the aforementioned electronic devices. For example, the electronic device may also include components that are more... Figure 4 The more or fewer components shown (such as network interfaces, display devices, etc.), or having the same Figure 4 The different configurations shown.

[0097] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing the hardware related to the terminal device. The program can be stored in a computer-readable storage medium, which may include: flash drive, ROM, RAM, disk or optical disk, etc.

[0098] According to another aspect of the present invention, a computer-readable storage medium is also provided, wherein a computer program is stored therein, wherein the computer program is executed by a processor to perform the steps in the above-described video push method.

[0099] Optionally, in this embodiment, those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing the hardware related to the terminal device. The program can be stored in a computer-readable storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0100] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0101] If the integrated units in the above embodiments are implemented as software functional units and sold or used as independent products, they can be stored in the aforementioned computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause one or more computer devices (which may be personal computers, servers, or network devices, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention.

[0102] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0103] In the several embodiments provided in this application, it should be understood that the disclosed client can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, indirect coupling or communication connection between units or modules, and may be electrical or other forms.

[0104] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0105] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0106] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A video push method, characterized in that, include: During the playback of historical videos, the facial expression information of the target object watching the historical videos, the target object's attention to the historical videos, and the interaction information between the target object and the historical videos are obtained. Determining a target video based on the facial expression information, the attention information, and the interaction information includes: determining the target object's interest score in the historical video based on a first score of the facial expression information, a second score of the attention information, and a third score of the interaction information; determining the target video based on the interest score; wherein the interaction information includes the target object's correctness data, completion time, and number of submissions for the related questions in the historical video; the third score is a value between 0 and 1, used to represent the target object's acceptability of the historical video; the correctness data is positively correlated with the third score; the completion time is negatively correlated with the third score; and the number of submissions is negatively correlated with the third score. The target video is pushed to the target object, wherein the target video is a video of the same or similar type as the historical video with the interest score exceeding the first threshold or the highest interest score; The interest score is determined by the weighted sum of the first score, the second score, and the third score, wherein the weight of the first score is greater than the weight of the second score and the weight of the third score.

2. The method according to claim 1, characterized in that, The step of determining the target object's interest score in the historical video based on the facial expression information, the attention information, and the interaction information further includes: A first score for the facial expression information, a second score for the attention information, and a third score for the interaction information are determined.

3. The method according to claim 2, characterized in that, The facial expression information includes facial images of the target object during the viewing of the historical video, and determining a first score for the facial expression information includes: The facial image is input into an expression recognition model, which then identifies the emotion type of the target object. Count the number of times each emotion type occurs in the target group; The ratio of the number of times a positive emotion occurs to the total number of emotions in the emotion type is defined as the first score.

4. The method according to claim 2, characterized in that, The attention information includes the number of screen-switching operations performed by the target object, the number of skip operations performed by the target object, the number of unattended screens of the target object, and the total duration of the historical videos. The second score for determining the attention information includes: The number of screen switching operations, the number of skipped operations, the number of unattended screens, and the total duration are substituted into the first score formula to obtain the second score. The first score formula is used to calculate the second score, and the second score is negatively correlated with the total duration.

5. The method according to claim 1, characterized in that, Determining the target video based on the interest score includes: The first video is identified as the target video, wherein the first video is a video of the same type as historical videos with an interest score higher than a first threshold; or The second video is identified as the target video, wherein the second video is a video of a similar type to historical videos with an interest score higher than a first threshold; or The third video is identified as the target video, wherein the third video is a video of the same type as the historical video with the highest interest score; or The fourth video is identified as the target video, wherein the fourth video is a video of a similar type to the historical video with the highest interest score.

6. The method according to claim 1, characterized in that, The step of pushing the target video to the target object includes: The target video is pushed to the target object so that the target object can start playing the target video after finishing playing the current video; or The target video is pushed to the target object, and the playback order of the target video is placed after the video currently being played by the target object.

7. The method according to claim 1, characterized in that, Determining the target video based on the interest score includes: When there are multiple target objects, the interest score of each target object to the historical video is determined to obtain an interest score matrix; In the interest score matrix, target objects with similar interest scores for the same historical video are grouped into one object group; The target video is determined from the videos that have been viewed by other objects in the object group and have a high interest score.

8. A video push device, characterized in that, include: The acquisition module is used to acquire facial expression information of the target object watching the historical video, the target object's attention to the historical video, and the interaction information between the target object and the historical video during the playback of the historical video. The determination module is used to determine a target video based on the facial expression information, the attention information, and the interaction information. This includes: determining the target object's interest score in the historical video based on a first score of the facial expression information, a second score of the attention information, and a third score of the interaction information; and determining the target video based on the interest score. The interaction information includes the target object's correct / incorrect answers to related questions in the historical video, completion time, and number of submissions. The third score is a value between 0 and 1, representing the target object's acceptability of the historical video. The correct / incorrect answers are positively correlated with the third score, the completion time is negatively correlated with the third score, and the number of submissions is negatively correlated with the third score. The push module is used to push the target video to the target object, wherein the target video is a video of the same or similar type as the historical video with the interest score exceeding the first threshold or the highest interest score; The interest score is obtained by weighted summation of the first score, the second score, and the third score, wherein the weight of the first score is greater than the weight of the second score and the weight of the third score.

9. A computer-readable storage medium storing a computer program, characterized in that, The computer program is executed by the processor to perform the method described in any one of claims 1 to 7.

10. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to execute the method described in any one of claims 1 to 7 through the computer program.

Images