A data display method, device, equipment and readable storage medium

Abstract

Embodiments of the present application disclose a data display method, device and equipment and readable storage medium, relating to image processing technology and natural language processing technology in the field of artificial intelligence, wherein the method comprises: obtaining feature data of an object, the feature data of the object comprising emotional features of the object; determining a target format and target display content based on the feature data of the object, the target format being used to indicate the layout of the target display content; and displaying the target display content in a target interface according to the target format. By using the embodiments of the present application, the accuracy of data pushing can be improved, and user experience can be improved.

Description

Technical Field

[0001] This application relates to the fields of image processing technology and natural language processing technology, and in particular to a data display method, apparatus, device and readable storage medium. Background Technology

[0002] With the development of the internet, terminal devices are widely used in daily life. These devices display various types of information, allowing users to learn about things. Currently, information push notifications typically involve users actively selecting preferred tags upon first logging into the application, and then data is pushed to the user based on these tags and displayed on the screen. However, because user preferences can change at any time, this data push method has low accuracy and a poor user experience.

[0003] Currently, user preferences are generally inferred based on the click-through rate of content displayed on the interface, and the pushed information content is adjusted accordingly. This method of obtaining user preferences is relatively simple, resulting in low accuracy of data push. Summary of the Invention

[0004] This application provides a data display method, apparatus, device, and readable storage medium, which can improve the accuracy of data push and enhance user experience.

[0005] In a first aspect, this application provides a data display method, including:

[0006] Obtain the object's feature data, which includes the object's emotional characteristics;

[0007] The target layout and target display content are determined based on the characteristic data of the object. The target layout is used to indicate the layout of the target display content.

[0008] Display the target content in the target interface according to the target layout.

[0009] In conjunction with the first aspect, in one possible implementation, the face image includes multiple objects;

[0010] The face image is subjected to face detection to determine the non-emotional features and emotional features of the object, including:

[0011] Face detection is performed on each object in the face image, and the confidence level of each object is determined. Based on the non-emotional features of the object with the highest confidence level and the emotional features of the object with the highest confidence level, the non-emotional features and the emotional features of the object are determined.

[0012] Secondly, this application provides a data display device, comprising:

[0013] The feature acquisition unit is used to acquire feature data of an object, including the object's emotional features.

[0014] The layout determination unit is used to determine the target layout and target display content based on the feature data of the object. The target layout is used to indicate the layout of the target display content.

[0015] The content display unit is used to display the target content in the target interface according to the target layout.

[0016] Thirdly, this application provides a computer device, including: a processor, a memory, and a network interface;

[0017] The processor is connected to a memory and a network interface. The network interface is used to provide data communication functions, the memory is used to store computer programs, and the processor is used to call the computer programs so that the computer device containing the processor can execute the data display method.

[0018] Fourthly, this application provides a computer-readable storage medium storing a computer program adapted to be loaded and executed by a processor, so that a computer device having the processor performs the above-described data display method.

[0019] Fifthly, this application provides a computer program product or computer program that includes computer instructions that, when executed by a processor, implement the above-described data display method.

[0020] In this embodiment, since the object's feature data includes the object's emotional characteristics, it can reflect the object's current emotional state. Furthermore, a target layout and target display content are determined based on the object's feature data. The target layout indicates the arrangement of the target display content, and the target content is displayed on the target interface according to the target layout. Because the target layout and target display content are determined based on the object's emotional characteristics, and the object's emotions reflect its current mood, determining the target layout and target display content based on the object's emotions is targeted, more in line with the object's emotions, and therefore more in line with the object's preferences. By displaying the target content on the target interface according to the target layout, targeted data display can be achieved, improving the accuracy of data push and enhancing the user experience. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a schematic diagram of the architecture of a data processing system provided in an embodiment of this application;

[0023] Figure 2 This is a schematic diagram illustrating an application scenario of a data display method provided in an embodiment of this application;

[0024] Figure 3 This is a flowchart illustrating a data display method provided in an embodiment of this application;

[0025] Figure 4 This is a schematic diagram of facial attribute information provided in an embodiment of this application;

[0026] Figure 5 This is a schematic diagram of facial landmark localization provided in an embodiment of this application;

[0027] Figure 6 This is a schematic diagram of the center point of the eyeball provided in an embodiment of this application;

[0028] Figure 7 This is a schematic diagram illustrating a display of content according to a target layout, provided in an embodiment of this application.

[0029] Figure 8 This is a flowchart illustrating a method for obtaining the theme and sentiment tags of initially displayed content, as provided in an embodiment of this application.

[0030] Figure 9 This is a schematic diagram of a neural network structure provided in an embodiment of this application;

[0031] Figure 10 This is a schematic diagram of the structure of a text detection model provided in an embodiment of this application;

[0032] Figure 11 This is a flowchart illustrating another data display method provided in an embodiment of this application;

[0033] Figure 12 This is a schematic diagram of the composition structure of a data display device provided in an embodiment of this application;

[0034] Figure 13 This is a schematic diagram of the composition structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0035] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0036] Artificial intelligence (AI) is a comprehensive discipline encompassing a wide range of fields, including both hardware and software technologies. Fundamental AI technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interactive systems, and mechatronics. AI software technologies primarily include computer vision, speech processing, natural language processing, and machine learning / deep learning.

[0037] Computer vision (CV) is a science that studies how to enable machines to "see." More specifically, it refers to machine vision, which uses cameras and computers to replace human eyes in recognizing and measuring targets, and then performs image processing to create images more suitable for human observation or transmission to instruments. As a scientific discipline, computer vision studies related theories and technologies, attempting to build artificial intelligence systems capable of extracting information from images or multidimensional data. Computer vision technologies typically include image processing, image recognition, image semantic understanding, image retrieval, OCR, video processing, video semantic understanding, video content / behavior recognition, 3D object reconstruction, 3D technology, virtual reality, augmented reality, simultaneous localization and mapping (SLAM), and common biometric recognition technologies such as facial recognition and fingerprint recognition.

[0038] In this application's embodiments, all user information-related data (such as object feature data) are data authorized by the user. This application relates to image processing and natural language processing technologies in the field of artificial intelligence. Optionally, for example, image processing technology can be used to obtain object feature data. Optionally, natural language processing technology can be used to obtain features of the initial displayed content in the target interface, such as the theme and sentiment tags of the initial displayed content. The technical solution of this application is applicable to scenarios where the content and target layout that an object is interested in are determined, and the content that the object is interested in is displayed according to the target layout, thereby achieving targeted information push. For example, when browsing an interface, by obtaining object feature data such as the object's emotional characteristics, the target layout and target display content can be determined based on the object's emotional characteristics, and the target display content can be displayed according to the target layout, which can improve the accuracy of data display, enrich data display methods, and enhance user experience.

[0039] Please see Figure 1 , Figure 1 This is a schematic diagram of the architecture of a data display system provided in an embodiment of this application, such as... Figure 1 As shown, the computer device can interact with the terminal device, and the number of terminal devices can be one or at least two. For example, when there are multiple terminal devices, the terminal devices may include... Figure 1 The system includes terminal devices 101a, 101b, and 101c. Taking terminal device 101a as an example, computer device 102 can acquire the characteristic data of the object. Further, computer device 102 can determine the target layout and target display content based on the object's characteristic data. Optionally, computer device 102 can also acquire the characteristics of the initial display content in the target interface, such as the theme and sentiment tags of the initial display content. Further, computer device 102 can combine the object's characteristic data, the theme of the initial display content, and the sentiment tags of the initial display content to determine the target layout and target display content. Further, computer device 102 can display the target display content in the target interface according to the target layout. Optionally, computer device 102 can also send the target layout and target display content to terminal device 101a so that terminal device 101a displays the target display content according to the target layout. Because the target layout and target display content that the object is interested in are determined by combining the object's characteristic data, the accuracy of data push can be improved, targeted data push can be achieved, and user click-through rates can be increased. Furthermore, it can enrich the data display methods and improve the user experience.

[0040] It is understood that the computer equipment mentioned in the embodiments of this application includes, but is not limited to, terminal devices or servers. In other words, the computer equipment can be a server or a terminal device, or a system composed of a server and a terminal device. The terminal device mentioned above can be an electronic device, including but not limited to mobile phones, tablets, desktop computers, laptops, PDAs, in-vehicle devices, intelligent voice interaction devices, augmented reality / virtual reality (AR / VR) devices, head-mounted displays, wearable devices, smart speakers, smart home appliances, aircraft, digital cameras, webcams, and other mobile internet devices (MIDs) with network access capabilities. The server mentioned above can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, vehicle-to-everything (V2X) communication, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0041] Further, please see Figure 2 , Figure 2 This is a schematic diagram illustrating an application scenario of a data display method provided in an embodiment of this application. For example... Figure 2 As shown, computer device 20 can acquire characteristic data of an object, such as the object's emotional characteristics. It can also acquire non-emotional characteristics of the object and browsing information about the object's initial display content on the display interface. Optionally, computer device 20 can acquire characteristics of the initial display content on the target interface, such as the theme and sentiment tag of the initial display content. Based on browsing information, the theme and sentiment tag of the initial display content, it can determine the content that the object is interested in. The content that the object is interested in may include the theme of the content (e.g., decompression) and the sentiment tag of the content (e.g., happiness). Further, computer device 20 can combine the content that the object is interested in with the object's non-emotional characteristics (e.g., age and gender) and emotional characteristics (e.g., happiness) to determine the target layout (dense, waterfall layout) and target display content. The target layout is used to indicate the arrangement of the target display content; thus, the target display content is displayed on the target interface according to the target layout. Optionally, computer device 20 can also acquire characteristic data of the object, such as the object's emotional characteristics, and determine the target layout and target display content based on the object's emotional characteristics.

[0042] Further, please see Figure 3 , Figure 3This is a flowchart illustrating a data display method provided in an embodiment of this application; as shown... Figure 3 As shown, this data display method can be applied to computer devices, and the data display method includes, but is not limited to, the following steps:

[0043] S101, Obtain the characteristic data of the object.

[0044] In this embodiment, the computer device can acquire feature data of an object, which may include the object's emotional characteristics. These emotional characteristics can indicate the user's current mood, such as surprise, happiness, neutrality, irritability, anger, sadness, etc. By determining the object's current emotional characteristics, the target layout and content that the object is interested in can be determined, thereby improving the user experience. Specifically, the object's emotional characteristics can be determined based on its facial features; for example, if the object's facial feature is laughing, it can indicate that the object's emotional characteristic is happiness.

[0045] Optionally, the object's feature data may also include the object's non-emotional features. The computer device can then combine the object's emotional and non-emotional features to determine the target layout and content that the object is interested in. The object's non-emotional features may include, but are not limited to, features such as the object's age group, gender, and interests. In other words, the target layout and content that the object is interested in can be determined based on the object's current emotion and its age group, gender, and interests. For example, if the object is a preschooler, it may be interested in animated content, and the layout it is interested in may be composed of anime elements, and so on.

[0046] Optionally, the object's feature data may also include browsing information about the object's initial display content on the target interface. This browsing information can reflect the object's liking of the initially displayed content. Browsing information may include viewing duration, browsing speed, etc. For example, if the viewing duration is greater than or equal to a duration threshold, it indicates that the user's liking of the initially displayed content is higher than or equal to the liking threshold, meaning the user likes the initially displayed content; if the viewing duration is less than the duration threshold, it indicates that the user's liking of the initially displayed content is lower than the liking threshold, meaning the user does not like the initially displayed content. Alternatively, if the browsing speed is greater than or equal to a speed threshold, it indicates that the user's liking of the initial displayed content is higher than or equal to the liking threshold; if the browsing speed is less than the speed threshold, it indicates that the user's liking of the initial displayed content is higher than or equal to the liking threshold. The target interface may refer to the current display interface of the computer device.

[0047] In other words, computer devices can determine the target layout and target display content by combining the object's emotional characteristics and the object's browsing information regarding the initially displayed content on the target interface, or by combining the object's emotional characteristics, the object's non-emotional characteristics, and the object's browsing information regarding the initially displayed content on the target interface. Optionally, computer devices can also determine the target layout and target display content based on the object's non-emotional characteristics; or based on the object's browsing information regarding the initially displayed content on the target interface; or by combining the object's non-emotional characteristics and the object's browsing information regarding the initially displayed content on the target interface.

[0048] Optionally, the computer device can acquire a facial image of an object and determine the object's feature data based on the facial image. Specifically, the computer device can acquire a facial image of an object, perform face detection on the facial image, and determine the object's non-emotional features and emotional features. Further, the computer device can detect the window scrolling speed of the target interface and determine the object's browsing information regarding the initially displayed content on the target interface based on the window scrolling speed.

[0049] Optionally, the computer device can monitor the window scrolling of the information stream (i.e., the target interface). When window scrolling is detected, the device can acquire the window scrolling position and scrolling speed to determine whether the object is sliding the displayed content on the target interface. When the object slides the displayed content on the target interface, if the sliding distance exceeds a sliding threshold, the camera device on the computer device, or a camera device associated with the computer device, can be activated to quickly capture a facial image.

[0050] In this context, "information stream" can refer to a list of published content arranged from top to bottom, such as news feeds in browsers, news articles, social media news lists, or Moments. When a facial image is acquired, computer devices can use artificial intelligence (AI) algorithms to perform face detection, determining the object's non-emotional and emotional characteristics. For example, facial attribute information can be obtained through a face detection interface, and non-emotional characteristics can be determined based on this information. Facial attribute information may include, but is not limited to, gender, age group, expression, and posture (e.g., whether wearing a hat or mask), and the corresponding non-emotional characteristics can include the object's gender, age group, etc. Optionally, emotional characteristics can be determined based on facial expressions in the facial attribute information. Specifically, a pre-defined correspondence between expressions and emotions can be established, such as a smiling expression corresponding to happiness, a blank expression corresponding to neutral emotions, etc. When facial attribute information is acquired, the emotion corresponding to the expression in the facial attribute information is determined as the object's emotional characteristic. Figure 4 As shown, Figure 4 This is a schematic diagram of facial attribute information provided in an embodiment of this application. The facial attribute information may include gender: male, expression: smiling, posture such as not wearing a hat, not wearing a mask, etc.

[0051] Optionally, facial landmark localization can be performed on the face image, and the emotional characteristics of the object can be determined by combining the facial landmark localization results with the facial expression information in the facial attribute information. Specifically, the computer device can acquire the first expression of the object in the facial attribute information corresponding to the face image; perform facial landmark localization on the face image to determine the facial landmarks in the face image, determine the second expression of the object based on the facial landmarks; and determine the emotional characteristics of the object based on the first and second expressions.

[0052] If the first and second expressions are the same, then the emotion corresponding to either the first or second expression can be identified as the object's emotional characteristic. For example, if both the first and second expressions are "happy," then "happy" can be identified as the object's emotional characteristic. If the first and second expressions are different, then the emotions corresponding to both the first and second expressions can be identified as the object's emotional characteristic. For example, if the first expression is "happy" and the second expression is "surprised," then both "surprised" and "happy" can be identified as the object's emotional characteristic. Alternatively, if the first and second expressions are different, then one of the emotions corresponding to the first and second expressions can be selected as the object's emotional characteristic.

[0053] Optionally, the computer device can combine AI algorithms or application programming interfaces (APIs) to obtain facial landmark information from a face image, and mark key locations on the face. For example, it can perform facial landmark localization on a face image, calculating 90 key points such as eyebrows, eyes, nose, mouth, facial contours, and pupils. Figure 5 As shown, Figure 5 This is a schematic diagram illustrating facial landmark localization provided in an embodiment of this application. Figure 5 In facial images, multiple facial features, such as eyebrows, eyes, nose, mouth, and pupils, correspond to multiple key points. By marking facial features with key points, multiple facial key points can be obtained. Furthermore, the marked point positions can be used as input to a model, and various custom facial expressions can be used as output. By training the model using a neural network, an expression detection model can be obtained. Based on this expression detection model, the correspondence between facial key points and facial expressions can be determined.

[0054] Optionally, the computer device can determine the object's interest characteristics in the displayed content based on the eye's gaze. If the target interface includes initial displayed content, the computer device can acquire eye region images from multiple consecutive frames of face images, and determine whether the object's gaze direction corresponds to the displayed content in the target display interface based on the position of the eyeball center point in the eye region images of the multiple face images; if so, the time difference value of the multiple face images can be acquired, thereby determining the browsing information for the displayed content based on the time difference value of the multiple face images. If it is determined that multiple face images all indicate that the object is viewing the displayed content, the time difference value between the last face image and the first face image in the multiple face images is acquired, and this time difference value is determined as the browsing information for the displayed content. Optionally, the eyeball center point in the eye region image can be as follows: Figure 6 As shown, Figure 6 This is a schematic diagram of the center point of an eyeball provided in an embodiment of this application, wherein one eye corresponds to one center point of the eyeball.

[0055] If the target interface includes multiple initial display contents, the computer device can acquire browsing information for each initial display content to determine the object's interest characteristics. Specifically, the computer device can acquire eye region images from multiple consecutive frames of facial images; determine the object's gaze direction based on the position of the eyeball center point in the eye region image; determine the display content corresponding to the gaze direction from the multiple initial display contents; determine the browsing information for each initial display content based on the time difference of the multiple frames of facial images; determine the object's liking for each initial display content based on the browsing information; and determine the object's interest characteristics based on the liking, the theme of each initial display content, and the sentiment tag of each initial display content.

[0056] In other words, by obtaining browsing information about each initially displayed content on the target interface, such as the viewing time or browsing speed of each initially displayed content, we can determine the object's liking for each initially displayed content. This allows us to determine the themes and sentiment tags of the content the object likes. When pushing data to the object in the future, we can combine the themes and sentiment tags of the content the object likes to improve the accuracy of data push and enhance the user experience.

[0057] Optionally, the computer device can acquire eye region images from a face image based on facial landmark information. By analyzing the relative relationship between the black center point of the eyeball and its outline, the device can determine the subject's gaze. Specifically, the computer device can pre-acquire a large number of face images showing the subject looking down, up, left, right, and looking directly. After marking the face images, the points in the eye region image are used as input to the model, and the results (looking down, up, left, right, and looking directly) are used as labels to train the gaze detection model. Thus, when acquiring the eye region image, the gaze detection model can determine the subject's gaze, which can include looking down, up, left, right, and looking directly, etc. By using the gaze detection model to detect the user's (i.e., the subject's) gaze and determine the subject's gaze direction, combined with the window scrolling speed of the target interface, the device can infer the subject's browsing information, such as viewing time, on each initially displayed content. Therefore, the viewing time can be used to determine the subject's liking for each initially displayed content, and the liking can be used to determine the subject's interest characteristics. For example, a longer viewing time indicates a higher level of liking for the initially displayed content; a shorter viewing time indicates a lower level of liking for the initially displayed content. This liking can be represented by a rating system or a numerical value, with higher ratings indicating higher liking and higher numerical values ​​indicating higher liking, and so on.

[0058] Optionally, the computer device can also acquire the browsing speed of the object under normal reading conditions. For example, it can determine the object's historical browsing speed based on its historical browsing history, and determine the object's liking for the initially displayed content based on the historical browsing speed and the object's browsing speed for the initially displayed content. If the object's browsing speed for the initially displayed content is greater than or equal to the historical browsing speed, then the object's liking for the initially displayed content is determined to be higher than or equal to a liking threshold. If the object's browsing speed for the initially displayed content is less than the historical browsing speed, then the object's liking for the initially displayed content is determined to be lower than a liking threshold.

[0059] Optionally, if the face image includes an object, the computer device can detect the face image of the object to determine the object's non-emotional features and emotional features.

[0060] Optionally, if the face image includes multiple objects, the computer device can determine the non-emotional features and emotional features of the objects by performing face detection on each object in the face image, determining the confidence level of each object, and determining the non-emotional features and emotional features of the objects based on the non-emotional features of the object with the highest confidence level and the emotional features of the object with the highest confidence level.

[0061] The confidence level can be determined based on the probability of each object in the face image representing each emotion. For example, if the face image includes three objects, the probability of object 1 being sad is 0.3, happy is 0.5, and surprised is 0.8. The probability of object 2 being sad is 0.61, happy is 0.4, and surprised is 0.39. The probability of object 3 being sad is 0.25, happy is 0.38, and surprised is 0.59. Therefore, 0.8 can be determined as the confidence level for object 1, 0.61 for object 2, and 0.59 for object 3. Furthermore, the non-emotional and emotional features of the object with the highest confidence level among the three objects are determined as the object's non-emotional and emotional features. Thus, the non-emotional features of object 1 are determined as the object's non-emotional features, and the emotional feature of object 1 (i.e., surprise) is determined as the object's emotional feature.

[0062] Optionally, confidence can also be determined based on the area occupied by the face of each object in the face image. The larger the area occupied by the face, the higher the confidence of the object; the smaller the area occupied by the face, the lower the confidence of the object. By identifying the non-emotional and emotional features of the object with the highest confidence from multiple objects, targeted data push and data display can be performed based on these non-emotional and emotional features, thereby improving user click-through rates.

[0063] Optionally, as shown in Table 1, the computer device can be pre-set with the correspondence between non-emotional features, emotional features, and viewing duration:

[0064] Table 1

[0065] Emotions (from positive to negative) gender age group browsing speed Duration of stay surprise male Toddlers Very slow First duration happy female juvenile Slower Second duration joy youth normal Third duration usually elderly Faster Fourth duration Irritability Very fast Fifth duration angry

[0066] In Table 1, browsing speed refers to the scrolling speed of the target interface window; the faster the scrolling speed, the shorter the dwell time; the slower the scrolling speed, the longer the dwell time. Dwell time reflects the duration an object spends viewing the initially displayed content on the target interface. Specifically, the first dwell time is greater than the second, the second is greater than the third, the third is greater than the fourth, and the fourth is greater than the fifth. It is understood that the categories of emotions are not limited to those in Table 1, and age groups, browsing speeds, and dwell times are not limited to those shown in Table 1; they can be adjusted according to needs.

[0067] S102, determine the target layout and target display content based on the object's feature data.

[0068] In this embodiment, the computer device can determine the target layout and target display content based on the object's characteristic data. The target layout indicates the arrangement and / or color scheme of the target display content. For example, when the target layout indicates the arrangement of the target display content, if the object's emotional characteristic is sadness, the target display content can be determined to be happy content, and the target layout can be a simple and relaxed mode. If the object's emotional characteristic is high stress, the target display content can be determined to be stress-relieving content, and the target layout can be a simple and relaxed mode. If the object's emotional characteristic is pleasure, the target display content can be determined to be happy content, and the target layout can be a dense waterfall layout, and so on. Since an object in a negative emotional state (such as irritability, high pressure, sadness) may not want to view too much display content, displaying content in a simple and relaxed layout can increase the probability of the object viewing the display content. When an object in a positive emotional state (such as pleasure, surprise), they are more likely to view multiple display contents; therefore, displaying content in a dense waterfall layout can improve the user experience. Alternatively, when the target layout is used to indicate the arrangement of the target content, if the target audience's emotional state is one of high stress, then the target content can be determined to be stress-relieving content, and the target color scheme could be refreshing, and so on. By combining the target audience's emotions with the color scheme for the target content, the color scheme can more easily adjust the audience's mood and improve the user experience.

[0069] Optionally, the target layout and target display content can be further determined by combining the displayed content in the target interface. Specifically, the computer device can obtain the theme and sentiment label of the initial displayed content in the target interface, and determine the target layout and target display content based on the object's feature data, the theme and sentiment label of the initial displayed content in the target interface.

[0070] In this embodiment, the computer device can obtain the theme and sentiment tag of the initially displayed content in the target interface. The sentiment tag can reflect the emotion of the displayed content, and may include, for example, surprise, happiness, pleasure, neutrality, annoyance, anger, etc. Sentiment tags can be divided into positive and negative tags; for example, the sentiment tags from surprise, happiness, pleasure, neutrality, annoyance, and anger sequentially change from positive tags to negative tags.

[0071] By obtaining the theme and sentiment tags of the initial displayed content in the target interface, and combining this with the user's browsing information regarding the initial displayed content, we can determine the user's liking for each piece of initial displayed content. This allows us to determine the user's liking for each theme and each sentiment tag, thus identifying the user's preferences. Subsequently, we can use these preferences to deliver targeted data, improving the accuracy of data delivery and enhancing the user experience.

[0072] Optionally, if the object's feature data includes any one of the object's non-emotional features, the object's emotional features, and the object's browsing information regarding the initially displayed content on the target interface, the computer device can combine the theme of the initially displayed content and the sentiment tag of the initially displayed content, as well as such feature data, to determine the target layout and the target display content.

[0073] If the object's characteristic data includes multiple aspects such as the object's non-emotional characteristics, the object's emotional characteristics, and the object's browsing information regarding the initially displayed content on the target interface, the computer device can combine the theme and sentiment tag of the initially displayed content with these multiple characteristic data to determine the target layout and target display content. The following explanation uses the example of object characteristic data including the object's non-emotional characteristics, the object's emotional characteristics, and the object's viewing time for the initially displayed content on the target interface:

[0074] If the object's feature data includes the object's non-emotional features, the object's emotional features, and the object's browsing information regarding the initially displayed content on the target interface, then the computer device can determine the object's interest features based on the browsing information, the theme of the initially displayed content, and the sentiment tag of the initially displayed content; and determine the target layout and target display content based on the object's interest features, the object's non-emotional features, and the object's emotional features.

[0075] In other words, by combining the object's non-emotional characteristics, the object's emotional characteristics, and the object's browsing information regarding the initially displayed content on the target interface, the object's preferences for the initially displayed content, as well as the object's favorite topic categories and sentiment tag categories, can be determined. Subsequently, display content containing the object's favorite topic categories and sentiment tag categories can be selected and pushed to the object, thereby improving the user experience and increasing the user click-through rate.

[0076] In some possible implementations, the target layout can be used to indicate the layout and / or color scheme of the target display content. The layout and / or color scheme of the target display content can be determined based on the type of characteristic data of the object or the type of the target display content, or it can be determined based on other methods. This application does not limit this.

[0077] For example, if the target layout is used to indicate the arrangement of the target displayed content, the computer device can determine the target layout based on the type of the object's characteristic data. If the object's characteristic data is of the first characteristic type, the target layout is determined to be a dense waterfall layout; if the object's characteristic data is of the second characteristic type, the target layout is determined to be a loose large heading layout; if the object's characteristic data is of the third characteristic type, the target layout is determined to be a simple and loose layout, and so on. The first characteristic type can include positive labels such as the object's emotion (surprise, happiness, joy), and browsing speed greater than a first speed threshold, etc. The second characteristic type can include neutral labels such as the object's emotion (normal), and browsing speed less than a second speed threshold, etc. The third characteristic type can include negative labels such as the object's emotion (anger, irritability), and browsing speed greater than a second speed threshold and less than a first speed threshold, etc.

[0078] The second speed threshold is less than the first speed threshold. Waterfall layout, also known as a waterfall flow layout, is a website page layout visually characterized by a staggered multi-column layout. As the page scrolls down, this layout continuously loads data blocks and appends them to the current end. "Dense" indicates that the initial amount of content displayed in the interface is greater than or equal to a certain threshold, while "loose" indicates that the initial amount of content displayed in the interface is less than a certain threshold. "Large Title" indicates that the font size of the title within the displayed content is larger than the font size of other content within the displayed content.

[0079] For example, if the target layout is used to indicate the color scheme of the target displayed content, the computer device can determine the target color scheme based on the type of the object's feature data. If the object's feature data is of the first feature type, the target color scheme is determined to be a bright mode. If the object's feature data is of the second feature type, the target color scheme is determined to be a high-contrast mode. If the object's feature data is of the third feature type, the target color scheme is determined to be a fresh mode, and so on. Here, a bright mode can refer to an overall bright tone and vibrant colors. A high-contrast mode can be used to indicate a large color difference between colors, such as a high contrast between black and white. The higher the contrast, the darker the dark parts of the image and the brighter the bright parts; the lower the contrast, the brighter the dark parts of the image and the darker the bright parts. A fresh mode can be used to indicate that the colors consist of light colors. The choice of color scheme can also be determined in conjunction with the gender of the object; for example, the same mode might correspond to light green, light pink, light blue, etc. Optionally, the color schemes for different genders can be different; for example, with a fresh color scheme, the color scheme for men could be light blue, and the color scheme for women could be light pink, etc.

[0080] For example, if the target layout is used to indicate the typography and color scheme of the target displayed content, the computer device can determine the target typography and target color scheme based on the type of the object's characteristic data. If the object's characteristic data is of the first characteristic type, then the target typography is determined to be a dense waterfall layout and the target color scheme to be a bright and cheerful style; if the object's characteristic data is of the second characteristic type, then the target typography is determined to be a relaxed, large-headline layout and the target color scheme to be a high-contrast style; if the object's characteristic data is of the third characteristic type, then the target typography is determined to be a simple and relaxed style and the target color scheme to be a clean and refreshing style, and so on.

[0081] For example, computer devices can also determine the target color scheme based on the type of content to be displayed. If the target content is of type one, the target color scheme is determined to be in the first color category. If the target content is of type two, the target color scheme is determined to be in the second color category, and so on. Here, type one could refer to stress relief or anxiety reduction, in which case the first color category could be colors like yellow or green, which can alleviate stress and anxiety. Type two could refer to eerie or suspenseful content, in which case the second color category could be colors like gray or black, which can create a sense of dread, and so on.

[0082] Optionally, if the target layout is used to indicate the typesetting and color scheme of the target display content, the method for determining the target layout and target display content based on the object's feature data may include: obtaining a target reference layout and a target reference color scheme that match the object's feature data from a correspondence table; determining the target reference layout as the typesetting of the target content, and determining the target reference color scheme as the color scheme of the target display content. The correspondence table can be used to store the correspondence between the object's feature data, reference layout, and reference color scheme.

[0083] Optionally, if the target layout is used to indicate the typography and color scheme of the target display content, the method for determining the target layout and target display content based on the object's feature data, the theme of the initial display content, and the sentiment tag of the initial display content may include: obtaining the target reference layout, target reference color scheme, and target reference theme that match the object's feature data from a correspondence table, wherein the correspondence table is used to store the correspondence between the object's feature data, reference layout, reference color scheme, and reference theme; determining the target reference layout as the layout of the target content, determining the target reference color scheme as the color scheme of the target display content, and determining the reference display content corresponding to the target reference theme as the target display content.

[0084] Specifically, the computer device can pre-set a correspondence table. This table reflects the correspondence between an object's characteristic data, reference layout, reference color scheme, and reference theme. When the object's characteristic data is determined, the target reference layout, target reference color scheme, and target reference theme corresponding to the characteristic data can be matched from the correspondence table to determine the target layout and target display content. For example, the correspondence table can be as shown in Table 2:

[0085] Table 2

[0086]

[0087] It is understandable that the reference layouts are not limited to those shown in Table 2, nor are the reference color schemes limited to those shown in Table 2.

[0088] Optionally, the computer device can arrange and combine various state dimensions to determine each layout and color scheme, with each state corresponding to a specific layout and color scheme. Optionally, the computer device can also use deep learning to perform controlled variable tests on a large number of objects, collect data on the emotions of a large number of objects, and the correspondence between these emotions and different layouts and color schemes, thereby obtaining a correspondence table. Based on this correspondence table, the target layout of the object can be quickly determined.

[0089] Optionally, the computer device can determine the theme and sentiment tags of the displayed content based on two dimensions: the theme and the object's characteristic data. For example, if the object's characteristic data includes "emotion: angry, gender: male, age group: young, browsing speed: normal," then the displayed content can correspond to themes and sentiment tags such as "psychology, stress relief, and anger." Simultaneously, the device can also consider the themes and sentiment tags of content that the object views for longer periods, allowing for targeted content display. If the object views content corresponding to a particular pair of themes for a longer period, then content corresponding to that theme can be prioritized. Optionally, the computer device can use word vector technology to obtain the spatial distance between the displayed content and the object's characteristic data (e.g., calculating the spatial distance between the theme of the displayed content and the object's sentiment tags, the spatial distance between the sentiment tags of the displayed content and the object's sentiment tags, the spatial distance between the theme of the displayed content and the object's interest features, etc.), and determine the set of content with the closest spatial distance as the target displayed content.

[0090] Optionally, the computer device can also obtain the object's historical layout information and determine the target layout information based on this information. The historical layout information may include layout information from the interfaces of all applications on the computer device. Alternatively, the historical layout information may include layout information from applications of the same type as the current application. For example, if the current application is a social application, the historical layout information may include layout information from the interfaces of other social applications on the computer device besides the current application. By obtaining the object's historical layout information, the object's preferences can be determined, and thus the target layout can be determined based on those preferences, improving the user experience.

[0091] Optionally, the computer device can also display icons for multiple layouts on the target interface, allowing the user to select the target layout from among several options. These layout icons may include, but are not limited to, the layout name and a thumbnail icon. By displaying icons for multiple layouts, users can easily choose their preferred layout, thereby improving the user experience.

[0092] Optionally, the computer device can adjust the target layout based on the emotional characteristics of the user. For example, if the user's emotional characteristic is the first emotional category, the target layout is determined to be the first layout; if the user's emotional characteristic is the second emotional category, the target layout is determined to be the second layout, and so on. Since users may have different preferences for the layout of the content displayed on the target interface depending on their emotional state, adjusting the target layout based on the user's emotional characteristics can enrich the data display methods and improve the user experience.

[0093] S103, Display the target content in the target interface according to the target layout.

[0094] In this embodiment, since the computer device obtains the target layout and target display content, it can display the target content in the target interface according to the target layout. The target layout may include, but is not limited to, the arrangement of the displayed content, the color scheme of the displayed content, the font size of the displayed content, the text type of the displayed content, and so on. By adjusting the layout in the interface in conjunction with the object's state, the layout can be made more in line with the object's mood, enriching the data display method and improving the user experience. For example, if the object is an elderly person who wears glasses, browses slowly, and is in a calm mood, a more concise and easy-to-read style (larger font, enhanced color contrast) can be displayed to match the content that the elderly person prefers to view when calm. Alternatively, if the object is a young person who is in a good mood and browses quickly, a more lively layout, such as a dense waterfall layout with a more cheerful color scheme, can be used to display the content, making the data display more interesting, in line with the object's characteristics, and achieving targeted data display.

[0095] like Figure 7 As shown, Figure 7This is a schematic diagram illustrating a content display according to a target layout, provided in an embodiment of this application. 7a represents the initial content layout, such as arranging content sequentially from top to bottom according to a fixed layout; 7b represents displaying content in a waterfall layout, such as displaying multiple pieces of content in a single line; 7c represents displaying content in a refreshing color scheme, such as adding refreshing colors to text backgrounds or displaying text in refreshing colors. For example, when the target's mood is pleasant, a more lively waterfall layout can be displayed, as shown in 7b. Alternatively, the theme color can be adjusted to better match the target's mood with the current layout and displayed content.

[0096] In this embodiment, since the object's feature data includes the object's emotional characteristics, it can reflect the object's current emotional state. Furthermore, a target layout and target display content are determined based on the object's feature data. The target layout indicates the arrangement of the target display content, and the target content is displayed on the target interface according to the target layout. Because the target layout and target display content are determined based on the object's emotional characteristics, and the object's emotions reflect its current mood, determining the target layout and target display content based on the object's emotions is targeted, more in line with the object's emotions, and therefore more in line with the object's preferences. By displaying the target content on the target interface according to the target layout, targeted data display can be achieved, improving the accuracy of data push and enhancing the user experience.

[0097] Optional, please see Figure 8 , Figure 8 This is a flowchart illustrating a method for obtaining the theme and sentiment tags of initially displayed content, as provided in an embodiment of this application. This method can be applied to computer devices; such as... Figure 8 As shown, the method includes, but is not limited to, the following steps:

[0098] S201, extract the text content from the initial display content, perform word segmentation on the text content, and obtain multiple words.

[0099] In this embodiment, the computer device can extract the text content of the initial displayed content in the target interface, perform word segmentation on the text content, and obtain multiple word segments. The text content can include text content in the displayed content and text content in the images within the displayed content. By performing word segmentation on the extracted text content, multiple word segments can be obtained, and keywords in the initial displayed content can be determined based on these multiple word segments. By extracting text from both the images and the initial displayed content, the accuracy of text extraction can be improved, as text in images often better reflects the meaning of the image. Optionally, the computer device can use Optical Character Recognition (OCR) methods to extract text from the images.

[0100] S202, perform clustering processing on multiple word segments to determine the keywords in the initial display content.

[0101] In this embodiment, since multiple word segments are extracted, the computer device can perform clustering processing on these multiple word segments to determine the keywords in the initial displayed content. The number of keywords can be one or more.

[0102] S203, determine the theme of the initial display content based on the keywords in the initial display content.

[0103] In this embodiment, if the number of keywords in the initially displayed content is one, then that keyword can be determined as the topic of the initially displayed content. If the number of keywords in the initially displayed content is multiple, then all of those keywords can be determined as the topic of the initially displayed content. Alternatively, the topic of the initially displayed content can be determined by combining all of those keywords.

[0104] Optionally, the computer device can use a keyword extraction algorithm to determine keywords in the initial displayed content, such as a Word2Vec word clustering keyword extraction algorithm. The computer device determines the word vectors of the multiple word segments, that is, converts the word segments into vector form. It then clusters the word segments in the initial displayed content using the K-Means algorithm, selects the cluster center as a primary keyword of the initial displayed content, calculates the distance between the word vectors of other word segments and the word vector of the cluster center (i.e., the similarity between vectors), and selects the K words closest to the cluster center as keywords, where K is a positive integer. The similarity between word segments can be calculated using vectors generated by Word2Vec.

[0105] In the specific implementation, the computer device can train a Word2Vec model on the corpus (i.e., the text content in the initial display content) to obtain a word vector file. Further, by preprocessing the text content, N candidate keywords are obtained, where N is a positive integer. The N candidate keywords are traversed, and their word vector representations are extracted from the word vector file. K-Means clustering is performed on the N candidate keywords to obtain the cluster centers for each category (the number of clusters can be preset). Further, the distance (Euclidean distance or Manhattan distance) between the word segments within each category and the cluster centers is calculated and sorted in descending order of cluster size. Further, the top K words with the highest sorted distances can be selected as keywords in the initial display content, thereby determining the topic of the initial display content based on these keywords. The third-party toolkit Scikit-learn provides functions related to the K-Means clustering algorithm: the `sklearn.cluster.KMeans()` function executes the K-Means algorithm, and the `sklearn.decomposition.PCA()` function is used for data dimensionality reduction.

[0106] S204, extract the image content from the initial display content, detect the image content, and determine the image sentiment tag corresponding to the image content.

[0107] In this embodiment, the computer device can extract image content from the initially displayed content, detect the image content, and determine the image sentiment tag corresponding to the image content. The image content includes the title image and other images in the initially displayed content.

[0108] Optionally, the computer device can use an image detection model to detect image content and determine the corresponding image sentiment tag. Specifically, by inputting the image content into the image detection model, the probability of the image content belonging to each of multiple sentiment categories can be determined based on the image detection model, and the sentiment category corresponding to the highest probability can be determined as the image sentiment tag corresponding to the image content. Alternatively, multiple sentiment categories with probabilities greater than a first threshold can be obtained from the probabilities of multiple sentiment categories, and all multiple sentiment categories can be determined as the image sentiment tags corresponding to the image content.

[0109] Optionally, when the computer device extracts image content from the initial display content, if the image content contains people and faces, it can use interfaces such as Tencent Cloud to analyze the facial expressions in the image content and determine the image's sentiment label based on the facial expressions. For example, when determining facial expressions, the sentiment label corresponding to the facial expression can be determined based on a predefined correspondence between facial expressions and emotions, serving as the image's sentiment label. If the image content does not contain people or faces, the computer device can use image detection models, such as neural networks like deep learning frameworks such as Convolutional Neural Networks (CNNs), to detect the image content and determine the corresponding sentiment label.

[0110] Optionally, before using a neural network to detect image content, the computer device can acquire a large number of images in advance, label the images (i.e., label the images with corresponding sentiment tags), and input the labeled images into the neural network for training. This allows the neural network to learn the correspondence between images and sentiment tags, and subsequently, the input images can be detected based on the neural network to determine the corresponding sentiment tags.

[0111] like Figure 9 As shown, Figure 9 This is a schematic diagram of a neural network structure provided in an embodiment of this application. The neural network includes a pre-trained weight structure, a pooling layer, and a transformation output layer. The pre-trained weight structure is used to pre-train the input sample data (i.e., the input image) to determine the features in the image. The pooling layer can reduce the size of the parameter matrix, thereby reducing the number of parameters in the final connected layers. In other words, the pooling layer can reduce the size of the parameter matrix in the feature map, reducing the number of parameters, which can speed up computation and prevent overfitting. The output layer can be used to transform the output result, representing the output result in the form of probabilities. By inputting an image into the neural network, the pre-trained weight structure processes the image to obtain the feature matrix corresponding to the features in the image. Further, processing the feature matrix based on the pooling layer can reduce its size. Finally, by transforming the features processed by the pooling layer through the output layer, the probability of the image belonging to each emotion category can be obtained.

[0112] S205, Detect the text content and determine the corresponding sentiment tags.

[0113] In this embodiment, the computer device can extract text from an image and merge it with the text content in the initial display content. That is, the text content can include both the text content in the initial display content and the text content in the image within the initial display content, thereby determining the corresponding sentiment tag. Optionally, the computer device can use a text detection model to detect the text content and determine the corresponding sentiment tag. Specifically, by inputting the text content into the text detection model, the probability of the text content belonging to each of multiple sentiment categories can be determined based on the text detection model, and the sentiment category with the highest probability can be determined as the corresponding sentiment tag. Alternatively, multiple sentiment categories with probabilities greater than a second threshold can be obtained from the probabilities of multiple sentiment categories, and all multiple sentiment categories can be determined as the corresponding sentiment tags for the text content.

[0114] Alternatively, computer devices can use WordVect word vector technology to extend text content to a high-dimensional space, and then use text detection models such as Recurrent Neural Networks (RNNs) or Transformer models to find the relationship between text content and sentiment. Improved RNN models, such as Long Short-Term Memory (LSTM) networks, can associate the context of a sentence, thus better identifying the relationship between text content and sentiment. Transformer models, represented by Self-attention, can perform parallel computation in a more efficient way to obtain the relationship between text content and sentiment labels.

[0115] like Figure 10 As shown, Figure 10This is a schematic diagram of the structure of a text detection model provided in an embodiment of this application. The model structure includes an input embedding layer, an encoder, a decoder, an output embedding layer, and a fully connected layer. The encoder includes a multi-head attention mechanism, a first normalization layer, a feedforward neural network layer, and a second normalization layer; the decoder includes an occlusion multi-head attention mechanism, a third normalization layer, a fourth normalization layer, a feedforward neural network layer, a fifth normalization layer, and a fully connected layer. In specific implementation, the computer device inputs text content into the model, and the input embedding layer converts the text content into vectors. By performing position encoding on the vectors, word position information can be added to the word vectors. Further, the position-encoded vectors are processed by the multi-head attention mechanism in the encoder to obtain the semantic features of the sentence; the semantic features are normalized based on the first normalization layer to obtain normalized semantic features; the normalized semantic features are mapped based on the feedforward neural network to obtain mapped features; and the mapped features are normalized based on the second normalization layer to obtain the encoder output. Furthermore, since the model includes multiple decoders, each with the same structure, the text features output by the previous decoder can be converted into vectors based on the output embedding layer. Positional encoding of these vectors yields positionally encoded feature vectors. Further, the positionally encoded feature vectors are processed using a masking multi-head attention mechanism in the decoder to obtain the semantic features of the sentence. The semantic features of the sentence are then normalized using a third normalization layer to obtain normalized semantic features. The encoder output and normalized semantic features are processed using a multi-head attention mechanism to obtain combined features. These combined features are then processed using a fourth normalization layer to obtain normalized combined features. The normalized combined features are then mapped using a feedforward neural network layer to obtain combined mapped features. The combined mapped features are then normalized using a fifth normalization layer to obtain normalized features. Finally, the normalized features are processed using a fully connected layer to obtain classification probabilities, i.e., the probability that the text content belongs to each sentiment category. Based on these probabilities, the relationship between the text content and the sentiment label is determined, thus obtaining the corresponding text sentiment label.

[0116] The input embedding layer transforms text content into vectors; the multi-head attention mechanism in the encoder captures semantic features between words in a sentence; the normalization layer normalizes the data, accelerating training and improving its stability. The feedforward neural network layer, through simple non-linear processing units, enables the model to perform non-linear processing, thus statically mapping multiple feature vectors output by the second normalization layer. The masking multi-head attention mechanism prevents the use of future output words during training. For example, during training, the first word cannot reference the generation result of the second word; masking sets this information to 0 to ensure that the information at prediction position i can only reference outputs smaller than i. Fully connected layers connect features.

[0117] S206, determine the sentiment tag of the initially displayed content based on the image sentiment tag and the text sentiment tag.

[0118] In this embodiment of the application, since the computer device obtains image sentiment tags and text sentiment tags, it can combine image sentiment tags and text sentiment tags to determine the sentiment tag of the initially displayed content.

[0119] In one possible scenario, if both image sentiment tags and text sentiment tags are of the same type, the computer device may determine both image sentiment tags and text sentiment tags as the sentiment tags for the initial displayed content; or it may select one of the image sentiment tags and text sentiment tags as the sentiment tag for the initial displayed content.

[0120] In another possible scenario, if there are multiple image sentiment tags and text sentiment tags, the computer device can obtain the weights corresponding to the image and the text, respectively, and determine the sentiment tag of the initially displayed content based on these weights. Specifically, the image sentiment tag indicates the first probability that the initially displayed content belongs to each of the multiple sentiment categories, and the text sentiment tag indicates the second probability that the initially displayed content belongs to each of the multiple sentiment categories; the computer device can obtain the first weight corresponding to the image content and the second weight corresponding to the text content, respectively; and determine the sentiment tag of the initially displayed content based on the first weight, the first probability, the second weight, and the second probability.

[0121] For example, if the emotion categories include four types: happy, surprised, neutral, and angry, and based on the image content, the probability of the initial displayed content being happy is 0.1, the probability of surprised is 0.1, the probability of neutral is 0.2, and the probability of angry is 0.6, and based on the text content, the probability of the initial displayed content being happy is 0.2, the probability of surprised is 0.2, the probability of neutral is 0.1, and the probability of angry is 0.5, and the first weight corresponding to the image content is obtained as 0.7, and the second weight corresponding to the text content is obtained as 0.3, then based on the first weight and the first probability... Based on the second weight and the second probability, the probability of the initial content being happy is: 0.1*0.7+0.2*0.3=0.13; the probability of the initial content being surprised is: 0.1*0.7+0.2*0.3=0.13; the probability of the initial content being normal is: 0.2*0.7+0.1*0.3=0.17; and the probability of the initial content being angry is: 0.6*0.7+0.5*0.3=0.57. The computer device can then determine the emotion category corresponding to the highest probability of 0.57, i.e., anger, as the final emotion label for the initial displayed content. Optionally, the computer device can also acquire multiple emotion categories with probabilities greater than a third threshold and determine these multiple emotion categories as the emotion labels for the initial displayed content.

[0122] In this embodiment, by detecting the images and text in the initial displayed content of the target interface, the theme and sentiment tags of the initial displayed content can be determined. Since the text extraction incorporates text from the images, and images often accurately reflect the meaning of the text, the accuracy of theme and sentiment tag determination can be improved. This, in turn, improves the accuracy of determining the target audience's preferences, subsequently enhancing the accuracy of data push notifications, and ultimately improving the accuracy of data display and user click-through rates.

[0123] Optional, please see Figure 11 , Figure 11 This is a flowchart illustrating another data display method provided in an embodiment of this application. This data display method can be applied to computer devices; such as... Figure 11 As shown, this data display method includes, but is not limited to, the following steps:

[0124] S301, Obtain the non-emotional characteristics of the object.

[0125] Among them, the non-emotional characteristics of the object are used to reflect the object's basic information, such as age group, hobbies, etc.

[0126] S302, Obtain the emotional characteristics of the object.

[0127] Among them, the emotional characteristics of an object can be used to reflect the object's current mood.

[0128] S303, Get the viewing duration of the object for the initial displayed content.

[0129] Among them, viewing time reflects the object's liking for the initially displayed content; the longer the viewing time, the higher the liking; the shorter the viewing time, the lower the liking.

[0130] S304, Get the theme of the initial display content.

[0131] The theme can reflect the meaning that the initially displayed content is intended to convey.

[0132] S305, Extract the image sentiment tags corresponding to the image content in the initial display content.

[0133] In this process, by extracting the image content from the initial display, the meaning of the image can be extracted, and then the image sentiment label can be determined based on the image meaning.

[0134] S306, Extract the text sentiment tags corresponding to the text content in the initially displayed content.

[0135] In particular, by extracting the text content of the initially displayed content, and including the text in the image, the text sentiment label determined based on the text content is more accurate.

[0136] S307, determine the sentiment tag of the initially displayed content based on image sentiment tags and text sentiment tags.

[0137] By combining image sentiment tags and text sentiment tags to determine the sentiment tag of the initial displayed content, the sentiment tag of the initial displayed content can be determined from different dimensions, thus improving the accuracy of data determination.

[0138] S308 determines the target display content and target layout based on non-emotional features, emotional features, viewing duration, theme, and sentiment tags.

[0139] The computer device can obtain the target layout from preset different state layouts. These preset state layouts can be obtained through deep learning, which involves controlling variables to test a large number of objects, collecting data on the emotions of these objects, and the correspondence between these emotions and different layouts and color combinations. Based on these preset state layouts, the target layout for an object can be quickly determined. Optionally, the preset state layouts can correspond to the layouts in the aforementioned correspondence table, allowing the target layout to be determined from the table.

[0140] S309, Display the target content according to the target layout.

[0141] Among these methods, displaying target content in the target interface according to a target layout is highly targeted because the target layout and target content are determined based on the user's current mood and their liking of the displayed content. This method can enrich data display methods, improve user experience, and increase click-through rates. The specific implementation methods of steps S301 to S309 in this embodiment can be found in [reference needed]. Figure 3 The corresponding steps S101 to S104, and Figure 8 The implementation methods of the corresponding steps S201 to S206 will not be described in detail here.

[0142] S310: Obtain the object's satisfaction level and adjust the preset different state layouts based on the satisfaction level.

[0143] Adjusting preset layouts for different states can refer to adjusting the correspondence between typesetting, object feature data, color matching, and themes. In this embodiment, satisfaction can be obtained by sending a satisfaction survey to the user and having them fill it out; or, the user's viewing time for the target content displayed on the target interface can be used to determine the user's satisfaction. Since user satisfaction reflects their liking for the content and layout displayed on the current target interface, the layouts in the preset layout library can be adjusted based on user satisfaction, such as adding, deleting, or modifying them. Optionally, the display content in the display content library can also be adjusted to improve user satisfaction. By continuously adjusting based on user feedback (satisfaction), the accuracy of data push can be improved, thereby improving user experience and ultimately increasing user click-through rate. Optionally, the method for determining the sentiment tag of the initial display content can also be adjusted based on user satisfaction, as can the method for obtaining the theme of the initial display content.

[0144] By analyzing factors such as the user's scrolling speed, viewing time, and emotional state when browsing newly changed content, we can further determine the user's satisfaction with the pushed content and target layout. This data serves as training data to optimize the accuracy of machine learning in labeling content and layout. In other words, we adjust the displayed content and target layout based on user satisfaction to increase click-through rates and improve user experience.

[0145] Optionally, the technical solution of this application can be applied to desktop and web terminals of computer devices. Through desktop and web terminals, page scrolling speed can be monitored, and front-facing camera captures images to obtain object feature data and the theme of the initially displayed content in the target interface, as well as the sentiment label of the initially displayed content. The dimensions in the technical solution of this application can also be increased or decreased, and more complex data collection can be achieved through more sensors. Optionally, the emotional characteristics of the object can be labeled emotions, or numerical representations between 0 and 1 (e.g., a larger value indicates greater happiness); or it can be a probability distribution of several emotions, which is not limited in this embodiment.

[0146] In this embodiment, since the object's feature data includes the object's emotional characteristics, it can reflect the object's current emotional state. Furthermore, a target layout and target display content are determined based on the object's feature data. The target layout indicates the arrangement of the target display content, and the target content is displayed on the target interface according to the target layout. Because the target layout and target display content are determined based on the object's emotional characteristics, and the object's emotions reflect its current mood, determining the target layout and target display content based on the object's emotions is targeted, more in line with the object's emotions, and therefore more in line with the object's preferences. By displaying the target content on the target interface according to the target layout, targeted data display can be achieved, improving the accuracy of data push and enhancing the user experience.

[0147] The methods of the embodiments of this application have been described above, and the apparatus of the embodiments of this application will be described below.

[0148] See Figure 12 , Figure 12 This is a schematic diagram of the structural composition of a data display device provided in an embodiment of this application. The data display device can be a computer program (including program code) running on a terminal device; the data display device can be used to execute corresponding steps in the data display method provided in the embodiment of this application. For example, the data display device 120 includes:

[0149] The feature acquisition unit 1201 is used to acquire feature data of an object, including the object's emotional features.

[0150] The layout determination unit 1202 is used to determine the target layout and target display content based on the feature data of the object, wherein the target layout is used to indicate the layout of the target display content;

[0151] The content display unit 1203 is used to display the target display content in the target interface according to the target layout.

[0152] Optionally, the data display device 120 further includes: a content acquisition unit 1204, used to acquire the theme of the initial display content in the target interface and the sentiment tag of the initial display content;

[0153] The layout determination unit 1202 is specifically used to determine the target layout and target display content based on the feature data of the object, the theme of the initial display content, and the sentiment tag of the initial display content.

[0154] Optionally, the target layout is also used to indicate the color scheme of the target displayed content; the layout determination unit 1202 is specifically used for:

[0155] Retrieve the target reference layout, target reference color scheme, and target reference theme that match the feature data of the object from the correspondence table. This correspondence table is used to store the correspondence between the object's feature data, reference layout, reference color scheme, and reference theme.

[0156] The target reference layout is determined as the layout of the target content, the target reference color scheme is determined as the color scheme of the target display content, and the reference display content corresponding to the target reference theme is determined as the target display content.

[0157] Optionally, the content acquisition unit 1204 is specifically used for:

[0158] Extract the text content from the initial displayed content, and perform word segmentation on the text content to obtain multiple words;

[0159] Cluster the multiple word segments to determine the keywords in the initial displayed content;

[0160] The theme of the initial displayed content is determined based on the keywords in the initial displayed content;

[0161] Extract the image content from the initial display content, detect the image content, and determine the image sentiment tag corresponding to the image content;

[0162] The text content is analyzed to determine the corresponding sentiment tag.

[0163] The sentiment tag of the initially displayed content is determined based on the image sentiment tag and the text sentiment tag.

[0164] Optionally, the image sentiment tag is used to indicate a first probability that the initially displayed content belongs to each of the multiple sentiment categories, and the text sentiment tag is used to indicate a second probability that the initially displayed content belongs to each of the multiple sentiment categories; the content acquisition unit 1204 is specifically used for:

[0165] Obtain the first weight corresponding to the image content and the second weight corresponding to the text content respectively;

[0166] Based on the first weight, the first probability, the second weight, and the second probability, the sentiment label of the initially displayed content is determined.

[0167] Optionally, the feature data of the object also includes the object's non-emotional features and the object's browsing information regarding the initially displayed content on the target interface; the layout determination unit 1202 is specifically used for:

[0168] Based on the browsing information, the theme of the initially displayed content, and the sentiment tag of the initially displayed content, determine the object's interest characteristics;

[0169] Based on the object's interest characteristics, non-emotional characteristics, and emotional characteristics, determine the target layout and the target display content.

[0170] Optionally, the feature acquisition unit 1201 is specifically used for:

[0171] Obtain the face image of the object, perform face detection on the face image, and determine the non-emotional features and emotional features of the object;

[0172] The window scrolling speed of the target interface is detected, and the browsing information of the object for the initially displayed content of the target interface is determined based on the window scrolling speed.

[0173] Optionally, the target interface includes multiple initial display contents; the feature acquisition unit 1201 is specifically used for:

[0174] Obtain eye region images from multiple consecutive frames of face images;

[0175] The object's gaze direction is determined based on the position of the eyeball's center point in the eye region image;

[0176] The display content corresponding to the gaze direction is determined from the multiple initial display contents, and the browsing information of each initial display content is determined based on the time difference of the multiple frames of face images;

[0177] The object's liking level for each initially displayed content is determined based on the browsing information of each initially displayed content. The object's interest characteristics are then determined based on the liking level, the theme of the initially displayed content, and the sentiment tag of the initially displayed content.

[0178] Optionally, the feature acquisition unit 1201 is specifically used for:

[0179] Obtain the facial attribute information corresponding to the face image, which includes the non-emotional features of the object and the object's first expression;

[0180] Facial landmarks are located in the face image to identify the key facial features, and the second expression of the subject is determined based on these key facial features.

[0181] The emotional characteristics of the object are determined based on the first expression and the second expression.

[0182] Optionally, the feature acquisition unit 1201 is specifically used for:

[0183] Face detection is performed on each object in the face image, and the confidence level of each object is determined. Based on the non-emotional features of the object with the highest confidence level and the emotional features of the object with the highest confidence level, the non-emotional features and the emotional features of the object are determined.

[0184] It should be noted that, Figure 12 For any content not mentioned in the corresponding embodiments, please refer to the description of the method embodiments, which will not be repeated here.

[0185] In this embodiment, since the object's feature data includes the object's emotional characteristics, it can reflect the object's current emotional state. Furthermore, a target layout and target display content are determined based on the object's feature data. The target layout indicates the arrangement of the target display content, and the target content is displayed on the target interface according to the target layout. Because the target layout and target display content are determined based on the object's emotional characteristics, and the object's emotions reflect its current mood, determining the target layout and target display content based on the object's emotions is targeted, more in line with the object's emotions, and therefore more in line with the object's preferences. By displaying the target content on the target interface according to the target layout, targeted data display can be achieved, improving the accuracy of data push and enhancing the user experience.

[0186] See Figure 13 , Figure 13 This is a schematic diagram of the structural composition of a computer device provided in an embodiment of this application. For example... Figure 13 As shown, the computer device 130 may include a processor 1301, a memory 1302, and a network interface 1303. The processor 1301 is connected to the memory 1302 and the network interface 1303, for example, the processor 1301 can be connected to the memory 1302 and the network interface 1303 via a bus. The computer device can be a terminal device or a server.

[0187] Processor 1301 is configured to support the image detection apparatus in performing the corresponding functions of the image detection method described above. Processor 1301 may be a Central Processing Unit (CPU), a Network Processor (NP), a hardware chip, or any combination thereof. The aforementioned hardware chip may be an Application-Specific Integrated Circuit (ASIC), a Programmable Logic Device (PLD), or a combination thereof. The aforementioned PLD may be a Complex Programmable Logic Device (CPLD), a Field-Programmable Gate Array (FPGA), a Generic Array Logic (GAL), or any combination thereof.

[0188] Memory 1302 is used to store program code, etc. Memory 1302 may include volatile memory (VM), such as random access memory (RAM); memory 1302 may also include non-volatile memory (NVM), such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid-state drive (SSD); memory 1302 may also include combinations of the above types of memory. In this embodiment of the invention, memory 1302 is used to store website security detection programs, interactive traffic data, etc.

[0189] Network interface 1303 is used to provide network communication functions.

[0190] Processor 1301 can call this program code to perform the following operations:

[0191] Obtain the feature data of the object, which includes the object's non-emotional features;

[0192] The target layout and target display content are determined based on the characteristic data of the object. The target layout is used to indicate the layout of the target display content.

[0193] Display the target content in the target interface according to the target layout.

[0194] It should be understood that the computer device 130 described in the embodiments of this application can perform the foregoing... Figure 3 , Figure 8 and Figure 11 The description of the data display method in the corresponding embodiments can also be performed as described above. Figure 12 The description of the data display device in the corresponding embodiments will not be repeated here. Furthermore, the beneficial effects of using the same method will also not be repeated.

[0195] This application also provides a computer-readable storage medium storing a computer program. The computer program includes program instructions that, when executed by a computer, cause the computer to perform the methods described in the foregoing embodiments. The computer can be part of the aforementioned computer device, such as the processor 1301 described above. As an example, the program instructions can be deployed on a single computer device for execution, or deployed on multiple computer devices located in one location for execution, or executed on multiple computer devices distributed across multiple locations and interconnected via a communication network. These multiple computer devices distributed across multiple locations and interconnected via a communication network can form a blockchain network.

[0196] This application also provides a computer program product or computer program that includes computer instructions. When executed by a processor, these computer instructions can implement some or all of the steps in the methods described above. Optionally, the computer instructions are stored in a computer-readable storage medium. The processor of a computer device can read the computer instructions from the computer-readable storage medium and execute the computer instructions, causing the computer device to perform the steps in the embodiments of the methods described above.

[0197] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0198] The above-disclosed embodiments are merely preferred embodiments of this application and should not be construed as limiting the scope of this application. Therefore, any equivalent variations made in accordance with the claims of this application shall still fall within the scope of this application.

Claims

1. A data display method, characterized in that, include: If it is detected that an object has performed a swipe operation on a target interface containing multiple initial display contents, and the swipe distance is greater than the swipe threshold, then the camera device is activated to capture a face image. Face detection is performed on the face image to obtain the feature data of the object. The feature data of the object includes the object's emotional features and non-emotional features. The non-emotional features are determined based on the facial attribute information corresponding to the face image. The facial attribute information includes posture information, which includes information indicating whether a hat or mask is being worn. If the face image includes multiple objects, the emotional features and non-emotional features of the objects are determined based on the probability of each object representing each emotion or the area occupied by the face, and the features of the object with the highest confidence are selected. Obtain the theme and sentiment tags of each of the multiple initially displayed contents in the target interface; The browsing information of the object for each initially displayed content is determined, and the object's liking for each initially displayed content is determined based on the browsing information of each initially displayed content. Then, the object's interest characteristics are determined by combining the liking, the theme and sentiment tags of the initially displayed content. Based on the object's interest characteristics, the object's non-emotional characteristics, and the object's emotional characteristics, a target layout and target display content are determined, wherein the target layout is used to indicate the arrangement of the target display content; The target display content is determined based on the spatial distance between the feature data of each display content and the object. The spatial distance is obtained based on word vector technology, and the spatial distance includes the spatial distance between the theme of the display content and the sentiment label of the object, and the spatial distance between the sentiment label of the display content and the sentiment label of the object. The target content is displayed in the target interface according to the target layout.

2. The method according to claim 1, characterized in that, The target layout is also used to indicate the color scheme of the target displayed content.

3. The method according to claim 1, characterized in that, The step of obtaining the theme and sentiment tags of each of the multiple initially displayed contents in the target interface includes: Extract the text content from the initial displayed content, and perform word segmentation on the text content to obtain multiple words; Clustering is performed on the multiple word segments to determine the keywords in the initial displayed content; The theme of the initial displayed content is determined based on the keywords in the initial displayed content; Extract the image content from the initial display content, detect the image content, and determine the image sentiment tag corresponding to the image content; The text content is detected to determine the corresponding sentiment tag. The emotion tag of the initially displayed content is determined based on the image emotion tag and the text emotion tag.

4. The method according to claim 3, characterized in that, The image sentiment tag is used to indicate the first probability that the initial displayed content belongs to each of the multiple sentiment categories, and the text sentiment tag is used to indicate the second probability that the initial displayed content belongs to each of the multiple sentiment categories; The step of determining the sentiment tag of the initially displayed content based on the image sentiment tag and the text sentiment tag includes: Obtain the first weight corresponding to the image content and the second weight corresponding to the text content respectively; Based on the first weight, the first probability, the second weight, and the second probability, the sentiment tag of the initial displayed content is determined.

5. The method according to claim 1, characterized in that, The determination of the browsing information of the object for each initially displayed content includes: The window scrolling speed of the target interface is detected, and the browsing information of the object for the initially displayed content in the target interface is determined based on the window scrolling speed.

6. The method according to claim 1, characterized in that, The target interface includes multiple initial display contents; The determination of the browsing information of the object for each initially displayed content includes: Obtain eye region images from multiple consecutive frames of face images; The object's gaze direction is determined based on the position of the eyeball's center point in the eye region image; The display content corresponding to the gaze direction is determined from the plurality of initial display contents, and the browsing information of each initial display content is determined based on the time difference of the multiple frames of face images.

7. The method according to claim 6, characterized in that, The step of performing face detection on the face image to obtain the feature data of the object includes: Obtain facial attribute information corresponding to the face image, wherein the facial attribute information includes the non-emotional features of the object and the object's first expression; Facial key points are located in the face image to determine the facial key points in the face image, and the second expression of the object is determined based on the facial key points; The emotional characteristics of the object are determined based on the first expression and the second expression.

8. A data display device, characterized in that, include: The feature acquisition unit is configured to, if it detects that an object has performed a sliding operation on a target interface containing multiple initial display contents, and the sliding distance is greater than a sliding threshold, activate a camera device to capture a face image; perform face detection on the face image to obtain feature data of the object, the feature data of the object including the object's emotional features and non-emotional features; wherein, the non-emotional features are determined based on facial attribute information corresponding to the face image, the facial attribute information including posture information, the posture information including information indicating whether a hat or mask is worn; and if the face image includes multiple objects, the emotional features and non-emotional features of the objects are determined based on the probability of each object representing each emotion or the area occupied by the face to determine the confidence level, and the features of the object with the highest confidence level are selected; The content acquisition unit is used to acquire the theme and sentiment tags of each of the multiple initially displayed contents in the target interface; determine the browsing information of the object for each initially displayed content, and determine the object's liking for each initially displayed content based on the browsing information of each initially displayed content, and then combine the liking, the theme and sentiment tags of the initially displayed content to determine the object's interest characteristics; The layout determination unit is used to determine a target layout and target display content based on the object's interest features, the object's non-emotional features, and the object's emotional features. The target layout is used to indicate the layout of the target display content. The target display content is determined based on the spatial distance between the feature data of each display content and the object. The spatial distance is obtained based on word vector technology, and the spatial distance includes the spatial distance between the topic of the display content and the object's emotion tag, and the spatial distance between the emotion tag of the display content and the object's emotion tag. The content display unit is used to display the target display content in the target interface according to the target layout.

9. A computer device, characterized in that, include: Processor, memory, and network interface; The processor is connected to the memory and the network interface, wherein the network interface is used to provide data communication functions, the memory is used to store program code, and the processor is used to call the program code so that the computer device executes the method according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program adapted to be loaded and executed by a processor to cause a computer device having the processor to perform the method of any one of claims 1-7.

11. A computer program product, characterized in that, The computer program product includes computer instructions that, when executed by a processor, implement the method as described in any one of claims 1-7.

Images