Information recommendation method, electronic equipment and computer readable storage medium

Abstract

The invention discloses an information recommendation method, electronic equipment and a computer readable storage medium. The method comprises the following steps: in response to a received first data request, based on environmental parameters and position transformation information of a target object, controlling expression acquisition equipment to acquire the target object to view a target expression image of current recommendation information; performing feature extraction on the target expression image to obtain expression features of multiple dimensions; analyzing the expression of the target object for viewing the current recommendation information based on the expression features of the multiple dimensions to obtain an expression analysis result; in response to a received second data request sent by the target client, obtaining object behavior data of the target object for the current recommendation information; and adjusting the current recommendation information based on the expression analysis result and the object behavior data to obtain target recommendation information. According to the invention, the technical problem of poor information recommendation effect in related technologies is solved.

Description

Technical Field

[0001] This invention relates to the field of information, and more specifically, to an information recommendation method, an electronic device, and a computer-readable storage medium. Background Technology

[0002] Currently, information search is crucial in areas such as advertising, game development, and intelligent customer service. Existing information search and recommendation systems have limitations in accuracy, real-time performance, adaptability to new users, and resistance to interference. Over-reliance on historical user data makes it difficult to capture in-depth and dynamic user needs. New users, lacking historical data, have limited personalized experiences, impacting search relevance and recommendation accuracy. Noisy data and abnormal behavior can easily interfere with the system, reducing judgment stability and resulting in poor information recommendation performance.

[0003] There is currently no effective solution to the above problems. Summary of the Invention

[0004] This invention provides an information recommendation method, an electronic device, and a computer-readable storage medium to at least address the technical problem of poor information recommendation performance in related technologies.

[0005] According to one aspect of the present invention, an information recommendation method is provided, comprising: responding to receiving a first data request, controlling an expression acquisition device to acquire a target expression image of the target object viewing current recommendation information based on environmental parameters and position change information of a target object, wherein the first data request is sent by the target object through a target client and is used to request the acquisition of the expression image corresponding to the target object; the environmental parameters are used to adjust the lighting of the expression acquisition device; and the position change information is used to adjust the image stabilization of the expression acquisition device according to the position change of the target object; extracting features from the target expression image to obtain multiple-dimensional expression features, wherein the different-dimensional expression features are used to represent the expression changes of the target object viewing the current recommendation information in different dimensions; and recommending an expression image of the target object viewing the current recommendation information based on the multiple-dimensional expression features. The system analyzes facial expressions to obtain facial expression analysis results, which quantify the target object's preference for the current recommendation information. In response to a second data request received from the target client, it acquires the target object's behavioral data related to the current recommendation information. This second data request is sent by the target object through the target client and requests the collection of corresponding behavioral data, representing the target object's interactive behavior data recorded in the target client over a historical time period. Based on the facial expression analysis results and the behavioral data, the current recommendation information is adjusted to obtain the target recommendation information, which is then sent to the target client for output. The target object's preference for the target recommendation information is greater than its preference for the current recommendation information.

[0006] Furthermore, based on environmental parameters and the target object's position change information, the facial expression acquisition device is controlled to acquire the target facial expression image of the target object viewing the current recommended information, including: determining the lighting adjustment parameters and anti-shake adjustment parameters based on environmental parameters and position change information; and controlling the facial expression acquisition device to acquire the target facial expression image of the target object viewing the current recommended information based on the lighting adjustment parameters and anti-shake adjustment parameters.

[0007] Furthermore, controlling the facial expression capture device to capture the target facial expression image of the target object viewing the current recommended information includes: detecting the current network environment of the facial expression capture device; and, if the network environment meets preset security conditions, controlling the facial expression capture device to capture the target facial expression image of the target object viewing the current recommended information.

[0008] Furthermore, the facial expressions of the target object viewing the current recommended information are analyzed based on multiple dimensions of facial expression features to obtain facial expression analysis results, including: analyzing the facial expressions of the target object viewing the current recommended information based on multiple dimensions of facial expression features to obtain initial facial expression analysis results; and desensitizing the initial facial expression analysis results based on preset key information to obtain facial expression analysis results, wherein desensitization processing is used to remove information in the initial facial expression analysis results that is associated with the preset key information.

[0009] Furthermore, the target recommendation information is sent to the target client so that the target recommendation information can be output in the target client, including: encrypting the target recommendation information to obtain encrypted data; sending the encrypted data to the target client so that the target recommendation information can be output in the target client, wherein the target recommendation information is obtained by the target client decrypting the encrypted data.

[0010] Furthermore, based on the facial expression analysis results and object behavior data, the current recommendation information is adjusted to obtain target recommendation information, including: fusing the facial expression analysis results and object behavior data to obtain fused data; generating an object model of the target object based on the fused data, wherein the object model is used to represent the target object's interests and preferences; and generating target recommendation information based on the object model and the target interaction scenario.

[0011] Furthermore, based on the object model and the target interaction scenario, target recommendation information is generated, including: generating data information and display information based on the object model and the target interaction scenario, wherein the data information is used to represent the data information that needs to be fed back to the target object, and the display information is used to represent the display method of the data information; and generating target recommendation information based on the data information and the display information.

[0012] Furthermore, the data information includes search information and recommendation information, and the display information includes search result display information and recommendation result display information; based on the object model and the target interaction scenario, the data information and display information are generated, including: if the target interaction scenario is a search scenario, generating search information and search result display information based on the object model; if the target interaction scenario is a recommendation scenario, generating recommendation information and recommendation result display information based on the object model.

[0013] According to another aspect of the present invention, an electronic device is also provided, comprising: a memory storing an executable program; and a processor for running the program, wherein the program executes the information recommendation method of the present invention during runtime.

[0014] According to another aspect of the present invention, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored executable program, wherein the information recommendation method of the present invention is executed in the executable program.

[0015] According to another aspect of the embodiments of this application, a computer program product is also provided, including a computer program that, when executed by a processor, implements the methods of various embodiments of this application.

[0016] According to another aspect of the embodiments of this application, a computer program product is also provided, including a non-volatile computer-readable storage medium storing a computer program, which, when executed by a processor, implements the methods in various embodiments of this application.

[0017] According to another aspect of the embodiments of this application, a computer program is also provided, which, when executed by a processor, implements the methods of the various embodiments of this application.

[0018] Through the above embodiments of this application, in response to receiving a first data request, based on environmental parameters and the target object's position change information, the facial expression acquisition device is controlled to acquire a target facial expression image of the target object viewing the current recommended information; feature extraction is performed on the target facial expression image to obtain multi-dimensional facial expression features; based on the multi-dimensional facial expression features, the facial expression of the target object viewing the current recommended information is analyzed to obtain facial expression analysis results; in response to receiving a second data request sent by the target client, object behavior data of the target object in response to the current recommended information is obtained; based on the facial expression analysis results and object behavior data, the current recommended information is adjusted to obtain target recommended information, and the target recommended information is sent to the target client so that the target recommended information can be output in the target client. By comparing the facial expression analysis results with the object behavior data, the gap between the target object's actual acceptance of the current recommended information and the expected emotional feedback is evaluated, and the recommendation strategy is dynamically improved using feedback. The above process is iterated repeatedly until the target recommended information and the target object's emotional state and behavioral preferences achieve a relatively good match, thereby improving the accuracy of the recommended content and user satisfaction, improving the information recommendation effect in related technologies, and thus solving the technical problem of poor information recommendation effect in related technologies. Attached Figure Description

[0019] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0020] Figure 1 This is a flowchart of an information recommendation method according to an embodiment of the present invention;

[0021] Figure 2 This is a flowchart of an optional information recommendation method according to an embodiment of the present invention;

[0022] Figure 3 This is a schematic diagram of an information recommendation device according to an embodiment of the present invention. Detailed Implementation

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

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

[0025] According to an embodiment of the present invention, an information recommendation method embodiment is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0026] Figure 1 This is a flowchart of an information recommendation method according to an embodiment of the present invention, such as... Figure 1 As shown, the method includes the following steps:

[0027] Step S102: In response to receiving the first data request, based on environmental parameters and the position change information of the target object, control the expression acquisition device to acquire the target expression image of the target object viewing the current recommended information.

[0028] The first data request is sent by the target object through the target client and is used to request the acquisition of the corresponding facial expression image of the target object. The environmental parameters are used to adjust the lighting of the facial expression acquisition device, and the position change information is used to adjust the anti-shake of the facial expression acquisition device according to the position change of the target object.

[0029] The aforementioned first data request can refer to a signal or request issued by the target object, typically sent to a server or other device over a network. In this case, the purpose of the request is to capture an image of the target object's facial expression. This request can be triggered by clicking a button, a voice command, or generated by the target client software or hardware.

[0030] The aforementioned environmental parameters can refer to external factors affecting image quality, such as light intensity and color temperature. In this embodiment, the system can monitor and intelligently adjust the lighting settings of the facial expression acquisition device in real time, ensuring that clear and accurate facial expression images can be obtained under different lighting conditions. The implementation of the lighting adjustment function improves the environmental adaptability of data acquisition and reduces the interference of environmental noise on facial expression recognition.

[0031] The target object mentioned above can refer to the user or entity that initiates the first data request. In this embodiment, the target object is an individual who wants to collect their facial expression images for purposes such as using a service, participating in an interaction, or performing user category verification.

[0032] The aforementioned positional change information can refer to the posture or positional changes of the target object while using the client. The system uses position sensors or image analysis technology to capture the positional changes of the target object in real time, and adjusts the anti-shake settings of the facial expression acquisition device accordingly to reduce image blurring caused by movement, thereby improving the clarity of facial expression images and the accuracy of facial expression recognition.

[0033] The aforementioned facial expression acquisition device refers to the component responsible for acquiring facial expression images of the target object in this embodiment of the application, which may be a camera and a micro-expression capture sensor. Upon receiving the first data request, the facial expression acquisition device is activated, acquiring changes in the target object's facial expressions, and simultaneously applying adaptive light adjustment and image stabilization technology to ensure that image quality is not affected by the environment.

[0034] The aforementioned current recommendation information can refer to personalized information or content suggestions provided by the system to the target user at a specific point in time, based on the target user's interests, behavioral history, and contextual information.

[0035] The aforementioned target facial expression images can refer to the image data of a user's facial expressions collected by the system in response to a first data request made by the target user through a target client (such as a mobile application, web interface, etc.). These images typically contain subtle changes in the user's facial muscles at a specific moment or in a specific context, reflecting the user's emotional state or emotional response to the information they have encountered.

[0036] The aforementioned light adjustment refers to the process by which the system dynamically adjusts the optical settings of the facial expression acquisition device based on environmental parameters (such as light intensity, direction, and color temperature) to ensure that the acquired image has high clarity and contrast.

[0037] The aforementioned anti-shake adjustment refers to automatically adjusting the stability settings of the facial expression acquisition device based on the positional changes of the target object, in order to reduce image blurring caused by movement, vibration, or unstable support.

[0038] The aforementioned target client can refer to the tool or platform used by the target object to send the first data request. The client can be a mobile application, computer software, web interface, or any application on a device capable of communicating with the server. In this scenario, the target client is responsible for receiving the user's instructions, converting them into a first data request that the system can understand, and then sending it out over the network.

[0039] In one optional embodiment, in response to a first data request sent by the target object through the target client, this embodiment activates the facial expression capture device and simultaneously analyzes environmental parameters and the target object's position change information to improve light adjustment and image stabilization. First, the device's exposure, sensitivity, and other settings can be automatically adjusted based on light intensity and color temperature to ensure image quality. Next, the target object's movement is monitored by a position sensor, and the image capture angle is corrected in real time to prevent image blurring and improve capture accuracy. Finally, the target facial expression image of the target object viewing the current recommended information is obtained.

[0040] This application embodiment, based on lighting conditions and the target object's movement, can stably and clearly capture micro-expression images of the target object when viewing recommended information. This improves the accuracy of subsequent micro-expression analysis and enhances the system's real-time ability to capture changes in user emotions, thereby enabling more accurate and faster adjustment of recommended content to meet users' personalized and immediate needs.

[0041] Step S104: Extract features from the target facial expression image to obtain facial expression features in multiple dimensions.

[0042] Among them, different dimensions of facial expression features are used to represent the changes in the target object's facial expressions when viewing the current recommended information in different dimensions.

[0043] The aforementioned feature extraction refers to the process of using computer vision technology to analyze target facial expression images. By identifying facial key points, expression regions, textures, and colors, a series of numerical features are extracted, including but not limited to the amplitude of eyebrow and eye movements, the degree of mouth curvature, and facial muscle tension. Feature extraction employs a multimodal fusion deep learning model, which can not only capture subtle changes in expression but also comprehensively consider the context of the expression to understand its complex meaning and potential emotional tendencies.

[0044] The aforementioned multi-dimensional facial expression features can encompass various types of attributes or descriptors. For example, in facial expression recognition, multiple features such as facial contours, eyebrow height, and mouth corner angles can be extracted from an image. Each feature reflects a different aspect of the expression, collectively forming a facial expression description.

[0045] The different dimensions of facial expression features mentioned above can refer to morphological features, such as changes in facial muscle shape, for example, frowning; dynamic features, such as the rate and duration of expression changes, for example, rapid blinking; local features, such as micro-expressions in specific areas of the face, for example, upturned corners of the mouth; overall features, such as the consistency of facial expressions, for example, a smiling face; emotional features, such as the basic emotions expressed, for example, happiness or sadness; and situational features, such as the adaptability of the expression to the situation, for example, laughing out loud while watching a comedy.

[0046] In one optional embodiment, firstly, computer vision technology can be used to locate key facial points and identify expression regions. Next, the dynamic changes in expressions are analyzed, recording dynamic features such as blink frequency and smile duration to capture emotional shifts. Then, focusing on local details such as the lips and corners of the eyes, local features such as the curvature of the corners of the mouth and wrinkles around the eyes are extracted to reveal deeper emotions. Simultaneously, the consistency of the overall facial expression is evaluated to form holistic features, ensuring a comprehensive understanding of the target audience's mindset. Based on this, emotion recognition algorithms are used, combined with contextual information, to analyze basic emotional features such as happiness and annoyance, further refining the emotion interpretation. Finally, considering the correlation between expressions and the environment, contextual features are analyzed to enhance the scene adaptability of expression analysis.

[0047] In another optional embodiment, the system first employs image processing technology to analyze the target facial expression image, extracting multi-dimensional facial features including morphological, dynamic, local, and overall characteristics. Morphological features quantify facial muscle changes, such as furrowed brows; dynamic features record the speed of expression changes, such as rapid blinking; local features focus on subtle movements of the corners of the mouth and eyes, such as an upturned corner of the mouth when smiling; and overall features assess the consistency of the entire facial expression to ensure a comprehensive interpretation. The system then combines contextual analysis to understand the specific emotions behind the expression.

[0048] This application's embodiments can acquire high-quality target facial expression images under any lighting conditions and with varying target object positions, and further extract multi-dimensional facial expression features. By extracting these multi-dimensional facial expression features, the system's ability to understand the target object's emotional state is significantly enhanced. Even without historical behavioral data, accurate recommendations can be made based on real-time facial expression changes, improving the matching degree between recommended content and user needs.

[0049] Step S106: Analyze the facial expressions of the target object when viewing the current recommended information based on multiple dimensions of facial expression features to obtain facial expression analysis results.

[0050] Among them, the facial expression analysis results are used to quantify the target audience's preference for the current recommended information.

[0051] The facial expression analysis results described above are used to represent the emotional state and potential needs of the target object in a specific interaction scenario. These results may include, but are not limited to, the target object's emotion category (such as happiness, sadness, surprise, etc.), emotion intensity (the strength of the emotion), and subconscious needs or interests.

[0052] In one optional embodiment, the first step is to capture an image of the target object's facial expression. Subsequently, multi-dimensional feature extraction is performed on the target expression image, which can encompass morphological, dynamic, local, and global expression features. Morphological features describe the static configuration of facial muscles, dynamic features analyze the rate of expression change, local features focus on subtle changes in key areas such as the eyes and corners of the mouth, and global features assess the consistency and intensity of the entire facial expression. These multi-dimensional features collectively reflect the target object's immediate emotional response to the recommended information.

[0053] Furthermore, embodiments of this application can employ deep learning algorithms to perform deep analysis and sentiment recognition on the extracted multi-dimensional facial expression features, producing facial expression analysis results. This process not only involves understanding the facial expression features but also requires effectively combining all collected multi-dimensional features to form a comprehensive and accurate quantitative indicator. The facial expression analysis results directly quantify the target audience's immediate preference for the current recommended information, thereby enabling a more nuanced capture of the target audience's emotional changes and attitudinal tendencies towards the recommended content.

[0054] Through multi-dimensional facial expression analysis, the system achieves multi-dimensional capture and interpretation of micro-expression features, transforming subtle changes in users' facial expressions into understandable and usable data. This not only enriches the information sources for user models but also significantly enhances their ability to capture real-time emotional changes, thereby accurately understanding and predicting users' immediate needs. In information search and recommendation scenarios, facial expression analysis results improve the personalization of search and recommendation content, ensuring that recommended information better matches users' emotional states and implicit preferences, effectively improving user satisfaction and the efficiency of information consumption.

[0055] Step S108: In response to receiving the second data request sent by the target client, obtain the target object's object behavior data in relation to the current recommendation information.

[0056] The second data request is sent by the target object through the target client and is used to request the collection of the target object's corresponding behavioral data. The object behavioral data is used to represent the target object's interactive behavior data during the historical time period recorded in the target client.

[0057] The aforementioned second data request can refer to an instruction issued by the target object through the target client to request the collection of its corresponding behavioral data. The second data request is typically triggered when a user uses a function associated with the embodiments of this application, indicating that the user agrees to share their behavioral data as a basis for generating personalized feedback information.

[0058] The aforementioned object behavior data can refer to all interactive behaviors of the target object recorded by the target client within a historical time period, including but not limited to clicks, browsing, searching, subscribing, favorites, ratings, and other operations.

[0059] The aforementioned target client refers to the electronic device application or platform with which the user interacts, capable of recording user behavior data and serving as a medium for communication between the system and the target object.

[0060] First, the target client detects that the user has triggered a second data request, indicating that the user has permitted the system to further collect their behavioral data to improve recommended content. In response, the system retrieves interaction records from the target client's database that are relevant to the target user and specific to the current recommendation. This retrieval process involves behavioral data across multiple dimensions, such as the number of clicks, dwell time, and feedback ratings of the target user on the recommendation, reflecting the user's reaction and attitude towards the recommended content. Subsequently, the system analyzes and processes the collected behavioral data to extract the user's true interests and behavioral patterns. By combining this with previous facial expression analysis results, the system can gain a deeper understanding of the user's emotional state and behavioral tendencies when receiving recommendation information, providing crucial information for generating highly personalized recommendations.

[0061] Upon receiving a request, the system can not only obtain real-time behavioral data of the target user in response to the recommended information, but also combine this data with behavioral records over historical time periods to construct a more comprehensive user behavior profile. Through the fusion analysis of facial expression analysis results and behavioral data, the system can intelligently identify users' emotional fluctuations and behavioral changes, dynamically adjust recommendation strategies, and provide information recommendations that better meet user needs.

[0062] The embodiments of this application can respond to the user's second data request, deeply analyze the behavioral data of the target object, and combine it with the expression analysis results to realize the dynamic adjustment of personalized recommendation content, enhance the intelligence and adaptability of the recommendation system, reduce recommendation errors, significantly improve the accuracy of information recommendation and user satisfaction, and solve the problem of poor recommendation effect in the prior art.

[0063] This embodiment of the application achieves behavioral data collection by acquiring historical behavioral data of the target object in response to a second data request. This enables real-time collection and updating of user behavioral data, ensuring that the recommendation system can dynamically adjust its recommendation strategy based on the latest behavior of the target object, thus improving the real-time nature and personalization of the recommendations. Secondly, by responding to the second data request, the system can selectively focus on specific behaviors of the target object related to the current recommendation information, rather than blindly collecting all behavioral data. This improves both the efficiency and quality of data collection and reduces interference from irrelevant data.

[0064] Step S110: Based on the facial expression analysis results and object behavior data, adjust the current recommendation information to obtain the target recommendation information, and send the target recommendation information to the target client so that the target recommendation information can be output in the target client.

[0065] Among them, the target object's preference for the target recommendation information is greater than the target object's preference for the current recommendation information.

[0066] The aforementioned object behavior data consists of a series of interactive behavior records of the target object in the target client, obtained through the second data request, revealing the target object's preferences and reactions to the recommended content.

[0067] The aforementioned current recommendation information refers to the recommendation content generated by the system based on existing information and provided to the target object before receiving the second data request.

[0068] The aforementioned target recommendation information is a new round of recommendations generated by the system by comprehensively analyzing facial expression results and object behavior data, and by adjusting the algorithm to better match the target object's preferences and emotional state.

[0069] The aforementioned preference level is a quantifiable indicator used to assess the target audience's interest in and satisfaction with the recommended content.

[0070] This application's embodiments first comprehensively evaluate the target object's preference for the current recommended information based on facial expression analysis results and object behavior data. This evaluation process involves complex emotion recognition and behavior analysis algorithms, the core of which lies in combining the target object's real-time emotional state with behavioral patterns to understand the target object's needs in a more comprehensive and in-depth way.

[0071] Subsequently, this embodiment employs an intelligent recommendation algorithm to adjust the recommendation strategy based on the evaluation results, generating target recommendation information. The adjustment strategy includes multiple aspects such as content type, theme, sentiment tendency, and presentation method, to ensure that the new recommendation information can more accurately match the target audience's preferences and improve the degree of preference. Finally, the target recommendation information is sent to the target client for output to the target audience. This output process also considers the target audience's emotional state and behavioral habits to provide a better user experience.

[0072] By combining real-time emotional and behavioral feedback from the target audience, personalized adjustments to the recommended content were achieved, ensuring that the target recommendations showed a significant increase in preference compared to the current recommendations. This process not only enhanced the accuracy of the recommendations but also demonstrated a meticulous attention to the user's individual needs.

[0073] In this embodiment, when a target user browses a news recommendation app, micro-expression data collection and analysis reveals a high level of interest in technology news. Subsequently, behavioral data indicates a tendency for the target user to read and share articles in the technology field in depth. Based on this information, the system adjusts its recommendation strategy, increasing the proportion of in-depth technology reports and high-quality articles. The resulting target recommendations are thus more aligned with the target user's interests, enhancing preference and providing a more satisfying information recommendation experience.

[0074] This application's embodiments can adjust the recommendation algorithm based on facial expression analysis results and object behavior data to generate targeted recommendation information. This process ensures that the recommended information better matches the preferences of the target object, improves the personalization and accuracy of the recommendations, and thus significantly enhances user satisfaction and the actual effectiveness of information recommendation.

[0075] In one alternative embodiment, the target feedback information may be embodied in a customized recommendation list, a personalized search result ranking, or an interface design adjusted according to the user's emotions, with the aim of providing services that are closer to the user's current intentions and needs.

[0076] The aforementioned preference level for recommended information refers to the user's level of liking or inclination towards the information content provided by the recommendation system. The preference level quantifies the user's acceptance and interest in specific recommended content, typically assessed by analyzing user feedback behaviors such as clicks, browsing time, likes, shares, or ignoring, to determine whether the user is satisfied or interested in the recommended information.

[0077] This application's embodiments achieve accuracy in recommended content by deeply analyzing the target object's real-time micro-expression and historical behavior data. This ensures that the target recommendation feedback information can more closely match the target object's preferences, effectively increasing their attention to and satisfaction with the recommended content, reducing invalid recommendations, and improving the attractiveness and conversion rate of the recommended content.

[0078] Through the above embodiments of this application, in response to receiving a first data request, based on environmental parameters and the target object's position change information, the facial expression acquisition device is controlled to acquire a target facial expression image of the target object viewing the current recommended information; feature extraction is performed on the target facial expression image to obtain multi-dimensional facial expression features; based on the multi-dimensional facial expression features, the facial expression of the target object viewing the current recommended information is analyzed to obtain facial expression analysis results; in response to receiving a second data request sent by the target client, object behavior data of the target object in response to the current recommended information is obtained; based on the facial expression analysis results and object behavior data, the current recommended information is adjusted to obtain target recommended information, and the target recommended information is sent to the target client so that the target recommended information can be output in the target client. By comparing the facial expression analysis results with the object behavior data, the gap between the target object's actual acceptance of the current recommended information and the expected emotional feedback is evaluated, and the recommendation strategy is dynamically improved using feedback. The above process is iterated repeatedly until the target recommended information and the target object's emotional state and behavioral preferences achieve a relatively good match, thereby improving the accuracy of the recommended content and user satisfaction, improving the information recommendation effect in related technologies, and thus solving the technical problem of poor information recommendation effect in related technologies.

[0079] Optionally, based on environmental parameters and the target object's position change information, the facial expression acquisition device is controlled to acquire the target facial expression image of the target object viewing the current recommended information, including: determining lighting adjustment parameters and anti-shake adjustment parameters based on environmental parameters and position change information; and controlling the facial expression acquisition device to acquire the target facial expression image of the target object viewing the current recommended information based on the lighting adjustment parameters and anti-shake adjustment parameters.

[0080] The aforementioned lighting adjustment parameters are calculated based on the lighting conditions information in the environmental parameters, including the required exposure settings, International Organization for Standardization Film Speed ​​Rating (ISO) sensitivity, aperture coefficient, and other parameters for the facial expression acquisition device, to ensure that the acquired target facial expression image is clear and has appropriate brightness.

[0081] The aforementioned image stabilization adjustment parameters are determined based on position change information to establish the image stabilization level or compensation value of the facial expression capture device, which is used to counteract the shaking of the target object or the device itself and ensure image quality.

[0082] In this embodiment, the first step is to determine lighting adjustment parameters and image stabilization adjustment parameters suitable for the current environmental conditions and the target object's posture, based on environmental parameters and position change information. Subsequently, the facial expression capture device is configured according to these parameters to capture facial expression images of the target object viewing the current recommended information in a better manner.

[0083] The system can dynamically adjust the parameters of the facial expression capture device to cope with complex operating environments and the movement of the target object. By analyzing environmental parameters in real time, such as light intensity and color temperature, the system can calculate lighting adjustment parameters, such as shutter speed, ISO value, and aperture size, ensuring that clear and appropriately bright images can be captured even in insufficient or excessively bright light conditions. Simultaneously, based on the positional changes of the target object, the system calculates image stabilization adjustment parameters, such as stabilization level and motion compensation algorithm, effectively suppressing image blur caused by the movement of the target object or the device itself, thus improving image quality.

[0084] For example, suppose the target subject is using a target client device (such as a news recommendation application) in an outdoor park. The system detects that the ambient light is strong, the color temperature is close to daylight, and the target subject is slightly moving. Based on these environmental parameters and positional changes, the system determines a lower ISO value, a faster shutter speed, and a smaller aperture value as lighting adjustment parameters to prevent overexposure. Simultaneously, based on the target subject's movement, a medium level of image stabilization is set as the image stabilization adjustment parameter to ensure image stability. The facial expression capture device adjusts accordingly, successfully capturing a series of clear and stable facial expression images, providing high-quality data support for subsequent expression analysis.

[0085] This application embodiment implements the function of dynamically adjusting the parameters of the facial expression acquisition device based on environmental parameters and location change information. This technical action ensures high-quality facial expression image acquisition, unaffected by changes in ambient lighting conditions or target object movement, providing a stable and reliable foundation for subsequent facial expression feature extraction and analysis. Adopting this dynamic parameter adjustment strategy not only improves the quality of facial expression image acquisition but also ensures a consistent recommendation experience for users when using the target client device in various environments. Even when the target object is under direct outdoor sunlight or in dim indoor lighting, the system can still ensure appropriate image clarity and brightness by adjusting lighting parameters; when the target object moves or the device is unstable, adjusting the anti-shake parameters ensures image stability and reduces blur and distortion. Optionally, controlling the facial expression acquisition device to acquire the target facial expression image of the target object viewing the current recommendation information includes: detecting the current network environment of the facial expression acquisition device; and, if the network environment meets preset security conditions, controlling the facial expression acquisition device to acquire the target facial expression image of the target object viewing the current recommendation information.

[0086] The aforementioned network environment refers to the network connection status of the facial expression acquisition device during operation, including but not limited to connection speed, stability, and security features, to ensure smooth data transmission and storage.

[0087] The aforementioned preset security conditions are predefined network security standards that ensure a sufficiently secure network environment when collecting user facial expression images, effectively protecting user data from unauthorized access.

[0088] Before acquiring target facial expression images, this embodiment first performs a comprehensive network environment check to assess whether it meets preset security conditions. This check covers multiple aspects, including the health of the network connection, data encryption capabilities, and the presence of potential threats. If the network environment is deemed secure, the system proceeds to the next step, controlling the facial expression acquisition device to begin acquiring facial expression images of the target object. Conversely, if the network environment is unsatisfactory, to protect user privacy and data security, the system will pause facial expression image acquisition until network conditions improve.

[0089] By detecting the network environment and setting strict preset security conditions before collecting target facial expression images, this application embodiment protects user privacy and data security. Even when using the target client device in uncertain network environments such as public networks, the system can intelligently determine the security of the environment and prevent data collection under insecure network conditions, thereby avoiding potential data leakage risks. Furthermore, this mechanism ensures high-quality transmission of facial expression image data, providing a foundation for subsequent facial expression recognition and analysis, and helping to improve the personalization and satisfaction of recommendation information.

[0090] The above steps enhance the level of user data protection, enabling the information recommendation system to securely and reliably utilize micro-expression data to improve recommendation effectiveness while respecting user privacy. Users no longer need to worry about their personal information being obtained in insecure network environments, increasing user trust in the system and improving the technological maturity and security of the entire recommendation system.

[0091] Optionally, the facial expressions of the target object viewing the current recommended information are analyzed based on multiple dimensions of facial expression features to obtain facial expression analysis results, including: analyzing the facial expressions of the target object viewing the current recommended information based on multiple dimensions of facial expression features to obtain initial facial expression analysis results; and desensitizing the initial facial expression analysis results based on preset key information to obtain facial expression analysis results, wherein desensitization processing is used to remove information associated with preset key information from the initial facial expression analysis results.

[0092] The aforementioned preset key information refers to specific information types or identifiers that are predefined and directly related to user privacy protection, such as the user's biometric features (e.g., facial contours) and fragments of personal information. This information needs to be concealed or replaced in subsequent processing to ensure that user privacy is protected to the maximum extent without affecting core functions.

[0093] The initial expression analysis results mentioned above refer to the raw, unmodified or masked output obtained directly after the system processes user micro-expression data using deep learning and sentiment analysis algorithms. The analysis results include not only the user's basic emotion category and intensity but also detailed information related to user characteristics. The initial expression analysis results retain the details extracted from the micro-expression data, allowing subsequent modules to utilize this rich information for deeper data fusion and user model construction.

[0094] The aforementioned desensitization process aims to remove or replace sensitive information contained in the original data to prevent unauthorized use or disclosure of this information and ensure the privacy and security of the data during the analysis and use phases.

[0095] In this embodiment, the system first performs in-depth analysis on the collected multi-dimensional facial expression features to interpret the subtle emotional changes of the target object when viewing the current recommended information, generating initial facial expression analysis results. Subsequently, the system desensitizes these initial results according to preset key information standards, removing all sensitive information that directly or indirectly reveals user characteristics, thereby obtaining the final facial expression analysis results.

[0096] In the deep analysis phase, the system utilizes deep learning neural networks, such as Convolutional Neural Networks (CNNs), to extract features and identify the target user's emotional category (such as joy, surprise, or disgust), intensity level, and duration when receiving recommendation information, forming multi-dimensional initial expression analysis results. This process accurately captures the user's emotional feedback to the recommended content, providing intuitive and valuable data support for adjusting the recommendation algorithm.

[0097] In the desensitization phase, the system compares the data against a pre-set list of key information to identify and remove information that could be linked to the user's personal identity. For example, through blurring, random replacement, or deletion, specific details of facial contours and precise pupil sizes are concealed, ensuring that no traceable personal characteristics are retained in the expression analysis results. By implementing expression feature analysis and targeted desensitization, this application can accurately analyze a user's emotional response to recommended information without infringing on user privacy.

[0098] In this embodiment, after generating the initial facial expression analysis results, measures are immediately taken to perform desensitization processing. For example, the system identifies the micro-expression features of a user while reading a recommended article, including subtle changes in the movement of the muscles around the eyes (indicating surprise or excitement), and an upward turn of the corners of the mouth (indicating happiness or satisfaction). However, in the analysis report, the system does not retain any specific numerical values ​​that can reconstruct the user's facial features. Instead, it converts these features into more general encoded forms, such as descriptive labels like "highly excited" or "extremely satisfied."

[0099] By employing a de-identification strategy, this application's embodiments ensure that personalized recommendations assisted by micro-expressions are implemented without disclosing user characteristic information, thus balancing technological innovation and privacy protection. Users can enjoy more personalized and accurate recommendation services while receiving privacy protection, increasing their acceptance and willingness to use the recommendation system.

[0100] Optionally, sending the target recommendation information to the target client so that the target recommendation information can be output in the target client includes: encrypting the target recommendation information to obtain encrypted data; sending the encrypted data to the target client so that the target recommendation information can be output in the target client, wherein the target recommendation information is obtained by the target client decrypting the encrypted data.

[0101] The encryption described above is a data protection method that uses a specific algorithm to convert target recommendation information into encrypted data to prevent it from being interpreted or tampered with by unauthorized parties during transmission.

[0102] The encrypted data mentioned above is the data form obtained after the target recommendation information has been processed by an encryption algorithm, which ensures the privacy and security of the information during transmission.

[0103] The decryption process described above is the process corresponding to encryption. The target client uses the corresponding decryption algorithm to restore the encrypted data into the target recommendation information so that the user can view and use it.

[0104] Before sending the target recommendation information to the target client, this embodiment first uses an encryption algorithm to encrypt the target recommendation information, generating encrypted data. Encryption ensures that the data exists in ciphertext form during transmission, preventing third-party decryption even if the data is intercepted, effectively protecting user privacy and data security. Subsequently, the encrypted data is sent to the target client through a secure data transmission channel. Upon receiving the encrypted data, the target client uses the corresponding decryption algorithm in its local environment to restore the encrypted data to the target recommendation information for user viewing and use. Both data transmission and decryption operations in this process are performed under strict security measures to prevent data leakage and unauthorized access.

[0105] By implementing encryption and decryption strategies, this application not only ensures the security of recommendation information during transmission but also protects user data privacy and improves the reliability of the entire recommendation system. The use of encrypted data avoids the plaintext display of recommendation information during transmission, reducing the risk of information leakage. Simultaneously, decryption is performed only on the target client, guaranteeing the user's exclusive access to the recommendation information. This mechanism achieves the dual goals of personalized and secure transmission of recommendation information, significantly enhancing user trust in the recommendation service and their willingness to use it.

[0106] For example, during the target user's use of the target client device, the system generates target recommendation information based on the user's behavioral data and facial expression analysis results. For instance, a technology news article. Before being transmitted to the target client, this recommendation information is encrypted using the Advanced Encryption Standard (AES) encryption algorithm, generating encrypted data. The encrypted data is sent to the target client through a secure network channel. After receiving the encrypted data, the target client uses a pre-configured AES decryption algorithm to restore the encrypted data to the original target recommendation information, i.e., the technology news article, for the user to read in a private and secure environment.

[0107] The above steps enhance the security and privacy protection capabilities of the information recommendation service. Users can use the target client device with peace of mind, reducing the risk of recommended content being intercepted during transmission. The encryption and decryption mechanisms not only protect the privacy of user data but also improve the security of recommended information transmission, serving as a crucial element of data protection in the recommendation system and demonstrating the emphasis and innovation of this application's embodiments in protecting user rights.

[0108] Optionally, based on the facial expression analysis results and object behavior data, the current recommendation information is adjusted to obtain target recommendation information, including: fusing the facial expression analysis results and object behavior data to obtain fused data; generating an object model of the target object based on the fused data, wherein the object model is used to represent the target object's interests and preferences; and generating target recommendation information based on the object model and the target interaction scenario.

[0109] The aforementioned fused data refers to combining facial expression analysis results with object behavior data to form a comprehensive data format that includes users' emotional states and behavioral patterns, used to more fully understand user needs.

[0110] The object model described above is used to represent a mathematical or logical model of the target object's interests and preferences. In the embodiments of this application, the object model is constructed based on fused data and can reflect the target object's immediate emotional feedback and long-term behavioral patterns in different contexts, providing a basis for personalized recommendations.

[0111] The target interaction scenarios described above are used to describe the specific context or environment in which the target object interacts with the application or platform, including but not limited to the user's current location, time, and type of device used, which helps the system better understand user needs and behaviors.

[0112] First, this embodiment of the application fuses facial expression analysis results and object behavior data. The data fusion process involves steps such as data preprocessing, feature extraction, and weight allocation to ensure the correlation and consistency of the two sets of data, generating fused data. Based on the fused data, the system uses machine learning algorithms or complex data analysis models to construct an object model of the target object. The object model comprehensively considers the real-time emotional fluctuations reflected in the user's micro-expressions and the long-term interests and preferences reflected in historical behavioral data, and can dynamically adjust to adapt to the latest needs of the target object. Finally, based on the constructed object model and the target interaction scenario, the system generates target recommendation information. The recommendation algorithm takes into account the user's current environment and state, such as a preference for sports news when exercising outdoors, and a greater focus on light entertainment content when resting at night, ensuring that the recommended information not only matches the user's interests but also matches the current context.

[0113] By implementing the above steps, the embodiments of this application can generate accurate and contextualized recommendation information based on more comprehensive user data. This object model construction method based on fused data effectively overcomes the limitations of traditional recommendation systems that rely solely on historical behavioral data while ignoring users' real-time emotional fluctuations, significantly improving the personalization and real-time responsiveness of recommendations. Furthermore, considering recommendation strategies tailored to target interaction scenarios, it can further enhance the relevance and attractiveness of recommendation information, improving user experience and satisfaction.

[0114] For example, when the target user is browsing information using a target client (a news reading application), the facial expression capture device captures the micro-expressions of interest displayed by the target user while reading a technology article. Simultaneously, the system records the user's behavior of searching and reading technology news multiple times over the past week. The system merges these two sets of data to construct an interest preference model for the target user, clearly indicating the target user's high interest in technology news. Considering that the target user is currently during a weekday lunch break, the system-generated recommendations not only include the latest technology articles but also specifically select easy-to-read technology trends and innovation stories relevant to the target user's recent career development, adapting to the target interaction scenario of relaxed reading during a lunch break. After receiving the recommendations, the user finds that the content not only matches their interests but also fits the current context, thereby improving the reading experience and user satisfaction with the embodiments of this application.

[0115] Through the above steps, the system can gain a more comprehensive understanding of user needs. The generated recommendations not only match user interests but can also be flexibly adjusted according to different interaction scenarios, ensuring high-quality recommended content and continuity of user interaction.

[0116] Optionally, based on the object model and the target interaction scenario, target recommendation information is generated, including: generating data information and display information based on the object model and the target interaction scenario, wherein the data information is used to represent the data information that needs to be fed back to the target object, and the display information is used to represent the display method of the data information; and generating target recommendation information based on the data information and the display information.

[0117] The aforementioned data refers to the specific data of the recommended content determined based on the object model of the target object and the target interaction scenario, including article titles, news summaries, product details, user ratings, etc., which are used to accurately convey information and value to the target object.

[0118] The aforementioned display information refers to the way data information is presented, including but not limited to font size, color, layout, icons, animation effects, etc., which aim to provide users with the most suitable visual experience based on the target interaction scenario.

[0119] This application first analyzes the target object's interests, emotional state, and behavioral patterns based on a constructed object model, while also considering the specificities of the target interaction scenario to generate data information. This process involves using machine learning algorithms to deeply mine user data to accurately match user needs. Subsequently, based on the characteristics of the target interaction scenario (e.g., the user's environment, time, device type, etc.), the system determines the configuration of the displayed information to provide the most comfortable reading or interaction experience. For example, when used at night, the system automatically adjusts the background color to dark mode; in outdoor bright light environments, it uses larger and brighter fonts and icons. Finally, combining the data information and display information, target recommendation information is generated to ensure that the information content accurately matches the user's interests, while the information presentation method meets the needs of the current interaction scenario, providing an improved user experience.

[0120] When a target user views recommended information on a target client (such as a personalized news app), the system constructs a current user model based on real-time micro-expression analysis and behavioral data. If the analysis shows that the target user exhibits a strong interest in technology news, and the target user is currently commuting in the morning (target interaction scenario), the system will generate a series of technology news-related data and display the information in a concise, fast-scrolling manner to suit the fast browsing habits of commuters, thus generating targeted recommended information.

[0121] By combining data with displayed information, the system can generate recommendations that align with user interests and the current interaction scenario, avoiding a disconnect between information content and display method, and ensuring effective information delivery and positive user interaction. This scenario-aware recommendation generation strategy improves the real-time nature and accuracy of information recommendations, enhancing user satisfaction.

[0122] For example, suppose a user uses a target client device (a personalized news application) during their morning commute. Based on the user's historical micro-expression analysis and behavioral data, the system learns that the user has a high interest in technology news. Simultaneously, considering the user's current commuting scenario, the system generates a series of data related to the latest technology trends, including article titles, summaries, and key images. To accommodate the need for quick browsing during commutes, the system uses concise display information, such as larger fonts, high-contrast colors, and a fast-scrolling article list. Ultimately, the user receives rich and easily browseable technology news recommendations on the target client, significantly improving information acquisition efficiency and reading experience during commutes.

[0123] Optionally, the data information includes search information and recommendation information, and the display information includes search result display information and recommendation result display information; based on the object model and the target interaction scenario, the data information and display information are generated, including: if the target interaction scenario is a search scenario, generating search information and search result display information based on the object model; if the target interaction scenario is a recommendation scenario, generating recommendation information and recommendation result display information based on the object model.

[0124] The search scenarios mentioned above refer to the specific environment, conditions, and context in which a user uses a search engine or recommendation system. Search scenarios encompass factors such as the time, location, purpose, device type, network status, previous search behavior, current activity level, and emotional state that influences search preferences.

[0125] The search information mentioned above refers to the keywords, topics, or categories generated by the system in a search scenario based on the target audience's interests and current needs, which are used to guide the target audience's search behavior.

[0126] The aforementioned recommendation scenarios can refer to the specific context and environment in which a user receives information or products pushed by the recommendation system, including but not limited to a collection of multi-dimensional information such as the user's physical location, time, device used, current activity, historical behavior patterns, personal preferences, social network status, and psychological or emotional state.

[0127] The aforementioned recommended information is a series of information or content automatically selected or generated by the system based on the object model of the target object in the recommendation scenario, aiming to meet the immediate interests and potential needs of the target object.

[0128] The search results display information described above describes how search results are presented, including the layout of the results list, font size, color, icons, etc., to ensure clear display and easy browsing of search results. The recommendation results display information described above is the presentation strategy for recommendation information, including the sorting of recommended content, highlighting, personalized tags, etc., aiming to provide an intuitive and attractive view of recommendation results for the target audience.

[0129] This application's embodiments address the needs of the target object in different interaction scenarios by employing flexible data generation and display strategies. First, the system determines whether the current scenario is a search or a recommendation scenario based on the target object's object model. For the search scenario, the system generates search information, including keywords or topics of interest to the target object, and designs search result display information to present the results in the most intuitive and efficient way, such as using larger fonts and a clear layout, enabling users to quickly locate content of interest in a fast-paced environment.

[0130] For recommendation scenarios, the system generates recommendation information based on the target object's object model, covering various types of content closely related to the target object's interests and preferences, such as technology news and movie recommendations. The design of the recommendation results display information focuses more on personalization and attractiveness. For example, based on the target object's micro-expression analysis results, the system decides to highlight content in the recommendation information that matches the user's emotional changes, and use dynamic icons or animation effects to attract the user's attention.

[0131] In search scenarios, such as when the target audience is using a target client (e.g., a news reading app) to search for information, the system generates search information based on the target audience's interest model, such as using "technology trends" as search keywords. At the same time, the system designs the search results display information to ensure that the results list is clear and easy to read, using high-contrast colors and bolded titles to make relevant technology news stand out in the list.

[0132] In recommendation scenarios, assuming the target audience is browsing a target client (such as a movie recommendation platform), the system automatically pushes a series of movie recommendations based on the target audience's object model, including recently popular science fiction films. The recommendation results display information including dynamic cover images and brief plot summaries to attract the target audience's attention. Simultaneously, the system adjusts the order of recommended content based on the target audience's micro-expression changes while watching trailers, placing the movies most likely to pique the user's interest at the top of the recommendation list.

[0133] This application's embodiments can provide users with customized data and display information based on different interaction scenarios, significantly improving the personalization of interaction and the comfort of the user experience. In search scenarios, the search information and search result display information generated by the system can quickly help users locate the information they need, improving search efficiency. In recommendation scenarios, the design of recommendation information and recommendation result display information focuses more on the attractiveness of the content and its matching with user interests, effectively improving the acceptance of recommended content and user satisfaction, and promoting long-term interaction between users and the system.

[0134] For example, suppose a user is searching based on their immediate interests within a target client (a personalized news reading app). The system, through micro-expression analysis, detects that the user shows a strong interest in technology news. At this point, the system generates search results containing "latest technological advancements" and designs the search results display with a clear layout and eye-catching headlines to ensure that technology news stands out among numerous search results. When browsing the search results, the user can quickly notice technology-related news headlines, improving search efficiency and reading experience.

[0135] Within the same application, when another user's target interaction scenario is a recommendation scenario, the system automatically generates a series of technology news recommendations matching the user's interests based on an object model. To attract the user's attention, the recommendation results display uses animated cover images and short summary reviews, making the recommended content appear more engaging on the user interface.

[0136] Figure 2 This is a flowchart of an information recommendation method according to an embodiment of this application, such as... Figure 2 As shown, the entire process involves five modules: micro-expression acquisition module, micro-expression analysis module, data fusion and user modeling module, search and recommendation generation module, and result display and interaction module. The role and main process of each module will be explained in detail below.

[0137] The micro-expression capture module employs an ultra-high-definition, high-speed professional camera and advanced micro-expression capture sensors to accurately capture subtle movements of the user's facial muscles and changes in facial expressions. It also features adaptive light adjustment and image stabilization technology to ensure high-quality micro-expression data acquisition in various environments. When the user opens a relevant application or service, the micro-expression capture device automatically starts; it captures micro-expression image data of the user's face at a high frame rate and accuracy; it performs preliminary screening and labeling of the captured data, removing obviously invalid data; it encrypts the captured data and transmits it to the micro-expression analysis module. Before collection, this module clearly explains the purpose, method, and scope of the collection to the user and obtains their explicit consent to protect user privacy. The collected data is encrypted and stored locally on the device, and a high-strength encryption algorithm is used during transmission to ensure data security and integrity.

[0138] The micro-expression analysis module utilizes a deep learning neural network model, trained on a large amount of labeled micro-expression data, to accurately interpret the emotions and subconscious information contained in micro-expressions. Combined with natural language processing technology, it transforms micro-expression features into quantifiable and understandable user demand indicators. It receives encrypted data from the micro-expression acquisition module and decrypts it in a secure environment; the data is then input into the deep learning model for feature extraction and sentiment analysis; detailed micro-expression analysis results are output, including emotion category, intensity, and potential demand tendencies. This module's analysis process takes place in an isolated, secure computing environment, with data access and processing limited to authorized personnel and algorithms. The analysis results are anonymized to remove sensitive information associated with individual categories.

[0139] The data fusion and user modeling module employs a unique fusion algorithm to deeply integrate micro-expression analysis results with users' traditional behavioral data (such as search history, browsing records, and click preferences). Utilizing a dynamic weighting mechanism, different weights are assigned based on the timeliness and importance of the data, constructing a comprehensive and real-time updated user model. This involves collecting and organizing users' traditional behavioral data; using the fusion algorithm to integrate micro-expression analysis results with traditional behavioral data; and building and updating the user model based on the fused data to reflect users' real-time interests and needs. During the data fusion process, this module employs encrypted computing technology to ensure data security. The user model is stored in encrypted form with access restrictions, and decryption is only performed during search and recommendation calculations.

[0140] The search and recommendation generation module, based on the fused user model, employs multimodal fusion algorithms, such as content-based multimodal fusion, collaborative filtering multimodal extension algorithms, and deep learning multimodal fusion algorithms, to generate accurate and diverse information search and recommendation results. According to the user model and the current search and recommendation context, multimodal fusion algorithms are used for calculation; a preliminary search and recommendation list is generated and ranked by relevance and priority. The search and recommendation results of this module are encrypted before being transmitted to the user, who decrypts and views them in a secure environment. The scope of use of the recommendation results is controlled to ensure they are not used for other unauthorized purposes.

[0141] The results display and interaction module employs an adaptive visual interface that dynamically adjusts the layout, colors, and font size based on the user's emotional state and preferences, providing a more comfortable and personalized interactive experience. It retrieves encrypted search and recommendation results and decrypts them on the user's end; displaying the results information to the user in an adaptive visual manner. This module protects user interaction data during results display and interaction by employing encrypted transmission and local storage technologies. Unnecessary interaction data is periodically deleted to ensure user privacy is protected.

[0142] According to another aspect of the present invention, an embodiment of an information recommendation method is also provided. It should be noted that the device can execute the information recommendation method of the above embodiment. The specific implementation method and preferred application scenarios are the same as those of the above embodiment, and will not be described again here.

[0143] Figure 3 This is a schematic diagram of an information recommendation device according to an embodiment of the present invention, such as... Figure 3 As shown, the device includes the following: acquisition module 302, feature extraction module 304, analysis module 306, acquisition module 308, and transmission module 310.

[0144] The system includes a data acquisition module, which, in response to a first data request, controls an expression acquisition device to acquire the target expression image of the target object viewing the current recommendation information based on environmental parameters and the target object's position change information. The first data request is sent by the target object through a target client and requests the acquisition of the target object's corresponding expression image. The environmental parameters are used to adjust the lighting on the expression acquisition device, and the position change information is used to adjust the image stabilization of the expression acquisition device according to the target object's position change. A feature extraction module is used to extract features from the target expression image, obtaining multi-dimensional expression features. These different dimensions of expression features represent the changes in the target object's expression when viewing the current recommendation information in different dimensions. An analysis module is used to analyze the target object's expression when viewing the current recommendation information based on the multi-dimensional expression features, obtaining... The system comprises: an expression analysis result, which quantifies the target object's preference for the current recommendation information; an acquisition module, which, in response to a second data request sent by the target client, acquires the target object's behavior data in response to the current recommendation information; the second data request, sent by the target object through the target client, requests the collection of the target object's corresponding behavior data, which represents the target object's interactive behavior data recorded in the target client over a historical time period; and a sending module, which, based on the expression analysis result and the behavior data, adjusts the current recommendation information to obtain the target recommendation information and sends it to the target client for output. The target object's preference for the target recommendation information is greater than its preference for the current recommendation information.

[0145] Optionally, the acquisition module is also used to determine the lighting adjustment parameters and the image stabilization adjustment parameters based on environmental parameters and position change information; and based on the lighting adjustment parameters and the image stabilization adjustment parameters, control the expression acquisition device to acquire the target expression image of the target object viewing the current recommended information.

[0146] Optionally, the acquisition module is also used to detect the current network environment of the expression acquisition device; if the network environment meets the preset security conditions, it controls the expression acquisition device to acquire the target expression image of the target object viewing the current recommended information.

[0147] Optionally, the analysis module is also used to analyze the facial expressions of the target object when viewing the current recommended information based on multiple dimensions of facial expression features to obtain initial facial expression analysis results; and to perform desensitization processing on the initial facial expression analysis results based on preset key information to obtain facial expression analysis results, wherein the desensitization processing is used to remove information in the initial facial expression analysis results that is associated with the preset key information.

[0148] Optionally, the sending module is also used to encrypt the target recommendation information to obtain encrypted data; and send the encrypted data to the target client so that the target recommendation information can be output in the target client, wherein the target recommendation information is obtained by the target client decrypting the encrypted data.

[0149] Optionally, the sending module is also used to fuse the facial expression analysis results and object behavior data to obtain fused data; generate an object model of the target object based on the fused data, wherein the object model is used to represent the target object's interests and preferences; and generate target recommendation information based on the object model and the target interaction scenario.

[0150] Optionally, the sending module is also used to generate data information and display information based on the object model and the target interaction scenario, wherein the data information is used to represent the data information that needs to be fed back to the target object, and the display information is used to represent the display method of the data information; and to generate target recommendation information based on the data information and display information.

[0151] Optionally, the sending module is also used to generate search information and search result display information based on the object model if the target interaction scenario is a search scenario; and to generate recommendation information and recommendation result display information based on the object model if the target interaction scenario is a recommendation scenario.

[0152] According to another aspect of the present invention, an electronic device is also provided, comprising: a memory storing an executable program; and a processor for running the program, wherein the program executes the methods of various embodiments of the present invention during runtime.

[0153] According to another aspect of the present invention, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored executable program, wherein, when the executable program is executed, it controls the device where the computer-readable storage medium is located to perform the methods of various embodiments of the present invention.

[0154] According to another aspect of the present invention, a computer program product is also provided, including a computer program that, when executed by a processor, implements the methods of various embodiments of the present invention.

[0155] According to another aspect of the present invention, a computer program product is also provided, including a non-volatile computer-readable storage medium storing a computer program that, when executed by a processor, implements the methods of various embodiments of the present invention.

[0156] According to another aspect of the present invention, a computer program is also provided, which, when executed by a processor, implements the methods of the various embodiments of the present invention.

[0157] Embodiments of this application also provide an electronic device, including: a memory storing an executable program; and a processor for running the program, wherein the program executes the methods in various embodiments of the present invention during runtime.

[0158] Embodiments of this application also provide a computer-readable storage medium including a stored executable program, wherein, when the executable program is executed, it controls the device where the computer-readable storage medium is located to perform the methods of various embodiments of the present invention.

[0159] Embodiments of this application also provide a computer program product, including a computer program that, when executed by a processor, implements the methods of various embodiments of the present invention.

[0160] Embodiments of this application also provide a computer program product, including a non-volatile computer-readable storage medium for storing a computer program that, when executed by a processor, implements the methods in various embodiments of the present invention.

[0161] Embodiments of this application also provide a computer program that, when executed by a processor, implements the methods described in the various embodiments of the present invention.

[0162] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0163] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some interfaces; indirect couplings or communication connections between units or modules may be electrical or other forms.

[0164] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0165] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0166] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0167] The above are merely preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. An information recommendation method characterized by comprising: include: In response to receiving a first data request, based on environmental parameters and the position change information of the target object, the facial expression acquisition device is controlled to acquire the target facial expression image of the target object viewing the current recommended information. The first data request is sent by the target object through the target client and is used to request the acquisition of the facial expression image corresponding to the target object. The environmental parameters are used to adjust the light of the facial expression acquisition device, and the position change information is used to adjust the anti-shake of the facial expression acquisition device according to the position change of the target object. Feature extraction is performed on the target facial expression image to obtain facial expression features in multiple dimensions. The facial expression features in different dimensions are used to represent the changes in the facial expression of the target object when viewing the current recommendation information in different dimensions. The facial expressions of the target object when viewing the current recommended information are analyzed based on the multiple dimensions of facial expression features to obtain facial expression analysis results, wherein the facial expression analysis results are used to quantify the target object's preference for the current recommended information; In response to receiving a second data request sent by the target client, object behavior data of the target object in relation to the current recommendation information is obtained, wherein the second data request is sent by the target object through the target client and is used to request the collection of corresponding behavior data of the target object, and the object behavior data is used to represent the interaction behavior data of the target object in the historical time period recorded in the target client; Based on the facial expression analysis results and the object behavior data, the current recommendation information is adjusted to obtain target recommendation information, which is then sent to the target client for output. The target object's preference for the target recommendation information is greater than its preference for the current recommendation information.

2. The method according to claim 1, characterized in that, Based on environmental parameters and the target object's position change information, the facial expression acquisition device is controlled to acquire the target facial expression image of the target object viewing the current recommended information, including: Based on the environmental parameters and the position change information, determine the light adjustment parameters and the image stabilization adjustment parameters; Based on the light adjustment parameters and the image stabilization adjustment parameters, the facial expression acquisition device is controlled to acquire the target facial expression image of the target object viewing the current recommended information.

3. The method according to claim 2, characterized in that, Controlling the facial expression acquisition device to acquire the target facial expression image of the target object viewing the current recommendation information includes: Detect the current network environment of the facial expression capture device; When the network environment meets preset security conditions, the facial expression acquisition device is controlled to acquire the target facial expression image of the target object viewing the current recommended information.

4. The method according to claim 1, characterized in that, Based on the multiple dimensions of facial expression features, the facial expressions of the target object when viewing the current recommendation information are analyzed to obtain facial expression analysis results, including: Based on the multiple dimensions of facial expression features, the facial expressions of the target object when viewing the current recommendation information are analyzed to obtain initial facial expression analysis results; Based on preset key information, the initial expression analysis result is desensitized to obtain the expression analysis result. The desensitization process is used to remove information in the initial expression analysis result that is associated with the preset key information.

5. The method according to claim 1, characterized in that, Sending the target recommendation information to the target client so that the target recommendation information can be output in the target client includes: The target recommendation information is encrypted to obtain encrypted data; The encrypted data is sent to the target client so that the target recommendation information is output in the target client, wherein the target recommendation information is obtained by the target client decrypting the encrypted data.

6. The method according to claim 1, characterized in that, Based on the facial expression analysis results and the object behavior data, the current recommendation information is adjusted to obtain target recommendation information, including: The facial expression analysis results and the object behavior data are fused to obtain fused data; An object model of the target object is generated based on the fused data, wherein the object model is used to represent the target object's interests and preferences; Based on the object model and the target interaction scenario, the target recommendation information is generated.

7. The method according to claim 6, characterized in that, Based on the object model and the target interaction scenario, the target recommendation information is generated, including: Based on the object model and the target interaction scenario, data information and display information are generated, wherein the data information is used to represent the data information that needs to be fed back to the target object, and the display information is used to represent the display method of the data information; Based on the data and the display information, the target recommendation information is generated.

8. The method according to claim 7, characterized in that, The data information includes search information and recommendation information, and the display information includes search result display information and recommendation result display information; Based on the object model and the target interaction scenario, data information and display information are generated, including: If the target interaction scenario is a search scenario, the search information and the search result display information are generated based on the object model; If the target interaction scenario is a recommendation scenario, the recommendation information and the recommendation result display information are generated based on the object model.

9. An electronic device, characterized in that, include: Memory, which stores executable programs; A processor for running the program, wherein the program, when running, performs the method according to any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored executable program, wherein, when the executable program is executed, it controls the device on which the storage medium is located to perform the method according to any one of claims 1 to 8.

Images