Adaptive information presentation method based on AR glasses and related equipment

Abstract

The invention discloses an information adaptive presentation method based on AR glasses and related equipment, and relates to the technical field of augmented reality. The technical means of evaluating the fatigue state of the user on the basis of cross-device use data, performing emotion recognition in combination with facial expression data, counting negative emotion feature events, weighting and calculating the content acceptance index on the basis of the negative degree and triggering information to simplify the display switching operation are adopted; therefore, when the screen use intensity of the user exceeds the personal historical average level, the device can actively monitor the emotional state change of the user, and the acceptance degree of the user to the current information content is evaluated by quantifying the occurrence frequency and duration of negative emotions. The complexity and density of information display are simplified when it is detected that the user generates a negative reaction, the problem of visual fatigue caused by the fact that AR glasses depend on a fixed watching duration triggering mechanism in the prior art is effectively solved, and then the technical effect of adjusting an information presentation strategy according to the real-time state of the user is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of augmented reality technology, in particular to an information adaptive presentation method based on AR glasses and related equipment. BACKGROUND

[0002] With the continuous maturity of augmented reality (AR) technology, AR glasses are gradually applied to complex scenes such as information display and retail.

[0003] In the related art, the AR glasses determine the fixation point of the user through the built-in eye movement tracking module or the head pose sensor. When the system detects that the user's line of sight stays on a certain identified object or area for more than a preset threshold time (i.e. fixation time), it will retrieve the information previously bound to the object or area from the database. These information is usually presented in the form of a standardized information panel containing multiple levels of menus, text details and three-dimensional models in a specific area of the user's field of view. The user can browse and interact layer by layer through eye selection, gestures or voice commands to obtain the content of their interest.

[0004] However, human attention is limited. The fixed fixation time trigger mechanism in the related art pops up a fixed information window, which may cause the AR glasses to continuously output high-density information for a long time, resulting in cognitive overload and visual fatigue for the user, reducing the information acquisition efficiency and the fluency of the user interaction experience. SUMMARY

[0005] The present application provides an information adaptive presentation method based on AR glasses and related equipment to improve the user's information acquisition efficiency.

[0006] In a first aspect, the present application provides an information adaptive presentation method based on AR glasses, applied to AR glasses, the method comprising: at the beginning of the current session of a target user, calculating the ratio of the total screen exposure time of the target user on the day to the average total screen exposure time in the preset historical period based on the obtained cross-device usage data, and taking the ratio as the diurnal accumulation index. The total screen exposure time is the cumulative value of the target AR glasses usage time and the total usage time of the associated terminal device associated with the target user; in the case where the diurnal accumulation index is determined to be higher than the preset fatigue threshold, obtaining a sequence of emotional feature events based on the facial micro-expression data of the target user in the preset time window combined with an emotion recognition algorithm; counting the number of occurrences and cumulative duration of negative emotional feature events in the sequence of emotional feature events, and calculating the content acceptance index based on the negative degree of the sequence of emotional feature events; in the case where the content acceptance index is determined to be lower than the preset negative reaction threshold, triggering an information simplification display switching operation.

[0007] By adopting the technical solutions, when the screen use intensity of the user exceeds the personal historical average level, the device can actively monitor the emotional state changes of the user, evaluate the acceptance degree of the user to the current information content by quantitatively analyzing the occurrence frequency and duration of the negative emotions, and automatically simplify the complexity and density of the information display when detecting the negative reaction of the user, thereby effectively solving the cognitive overload and visual fatigue problems caused by the fixed fixation time trigger mechanism of the AR glasses in the prior art, and achieving the technical effect of dynamically adjusting the information presentation strategy according to the real-time physiological and psychological state of the user.

[0008] In combination with some embodiments of the first aspect, in some embodiments, after the step of calculating the ratio of the total screen exposure time of the target user in the day to the average total screen exposure time in the preset historical period based on the obtained cross-device use data as the daily accumulation index at the beginning of the current session of the target user, the method further comprises: based on the actual sleep duration, the average sleep demand duration and the sleep quality score of the target user obtained by the associated terminal device, multiplying the ratio of the actual sleep duration to the average sleep demand duration by the sleep quality score to obtain a sleep recovery factor; and taking the ratio of the daily accumulation index to the sleep recovery factor as the updated daily accumulation index.

[0009] By adopting the technical solutions, the introduction of the sleep recovery factor can more accurately reflect the real fatigue state of the user. The ratio of the actual sleep duration to the average sleep demand duration reflects the sufficiency of the sleep of the user, and the sleep quality score reflects the effectiveness of the sleep. The sleep recovery factor obtained by multiplying the two reflects the physiological recovery ability of the user. The ratio calculation of the daily accumulation index and the sleep recovery factor realizes the dynamic correction of the accumulated state of the user fatigue, so that the fatigue evaluation not only considers the screen use intensity in the day, but also considers the recovery ability difference of the user. This personalized fatigue evaluation mechanism can more accurately identify the real fatigue level of the user, avoid underestimating the fatigue state caused by insufficient sleep, or ignoring the good recovery state caused by high-quality sleep, thereby improving the accuracy and effectiveness of the subsequent emotional monitoring and content optimization decisions.

[0010] In some embodiments in combination with the first aspect, in some embodiments, the step of statistically counting the number of occurrences and the cumulative duration of the negative emotional feature events within the emotional feature event, and calculating the content acceptance index based on the negative degree of the emotional feature event, specifically comprises: timestamp alignment of the emotional feature event sequence and the synchronous fixation point data obtained through eye tracking, binding one or more information content units that the target user fixated on at the corresponding time to each negative emotional feature event; calculating the ratio of the cumulative duration of the negative emotional feature event associated with the information content unit to the total duration of the target user's fixation to obtain the negative feedback rate of the information content unit; based on the information weight of the information content unit in the current information panel, the negative feedback rate is weighted and summed to obtain the content acceptance index.

[0011] By adopting the above technical solutions, since the emotional feature event sequence is timestamped and aligned with the eye tracking data to establish an emotion-content binding relationship, the negative feedback rate of the information content unit is calculated, and the weighted sum is performed based on the information weight, the system can identify the specific information content that triggers the user's negative emotion, and through quantitative analysis of the negative impact degree of different information units and comprehensive evaluation combined with their importance in the overall information, the technical problems of being unable to locate the source of negative emotion and lacking a targeted content optimization mechanism in the prior art are effectively solved, thereby realizing the technical effects of intelligent negative reaction recognition and personalized content acceptance evaluation based on emotion-content association.

[0012] In some embodiments in combination with the first aspect, in some embodiments, the step of triggering the information simplification display switching operation specifically comprises: performing semantic analysis on the currently displayed text content by calling a pre-trained natural language processing model to obtain a semantic analysis result, the semantic analysis including identifying information type labels and extracting key semantic entities, the information type labels including product description, technical specifications, and user feedback; based on the semantic analysis result, at least two candidate simplified information templates are generated according to a preset information reorganization rule, the simplified information template being a simplified information formed by retaining the content corresponding to a specific information type label and compressing the content corresponding to a non-specific information type label; based on the fixation duration of the target user on the at least two simplified information templates within a preset response time, the simplified information template with a fixation duration exceeding a preset selection threshold is determined as a target simplified information template; the information simplification display switching operation is performed based on the target simplified information template.

[0013] By adopting the technical solutions, the system can intelligently understand the semantic structure and type distribution of information content, generate targeted information simplification schemes for user selection, realize non-interference preference recognition through natural gaze behavior of the user, and perform accurate information simplification display operation, effectively solving the technical problems of single information simplification strategy and lack of user preference perception mechanism in the prior art, and achieving the technical effects of intelligent information reorganization based on semantic understanding and personalized simplification strategy selection based on natural interaction.

[0014] In some embodiments of the first aspect, based on the semantic analysis result, the step of generating at least two candidate simplified information templates according to a preset information reorganization rule specifically includes: after obtaining the negative feedback rates of all information content units, determining the information type label with the highest negative feedback rate among all information type labels as a high-negative-feedback information category and generating a first simplified information template, the first simplified information template being formed by deleting content belonging to the high-negative-feedback information category from the currently displayed page; based on the negative feedback rates, determining information categories with negative feedback rates lower than a preset safety threshold as a low-negative-feedback information category set, and generating a second simplified information template, the second simplified information template being formed by aggregating all content belonging to the low-negative-feedback information category set; and taking the first and second simplified information templates as the candidate simplified information templates.

[0015] By adopting the above technical solutions, the difference simplified strategy is generated by analyzing the negative feedback rate distribution, the first simplified information template removes the high-negative-feedback information category by using the deletion strategy, and the second simplified information template aggregates the low-negative-feedback information category by using the retention strategy, and the two strategies form a complementary candidate scheme set. The deletion strategy can quickly eliminate negative stimulus sources, and the retention strategy can maintain information integrity, and the user can select according to current needs and preferences. This dual-strategy parallel design fully considers the diversified needs of different users in different situations, meets the needs of users who require high information simplification, and also takes care of users who require high information integrity. By providing clear and differentiated selection schemes, the flexibility and adaptability of the simplification operation are improved, and the user can obtain an information presentation mode that meets the current state and task requirements.

[0016] In some embodiments in combination with the first aspect, after the step of determining the simplified information template with a gaze duration exceeding the preset selection threshold as the target simplified information template based on the gaze duration of the target user on the at least two simplified information templates within the preset response time, the method further comprises: in a case where it is determined that the target user does not respond within the preset response time, determining an intersection of information type labels in a historical interaction context and the current set of information type labels based on the information contained in the currently displayed information in a historical interaction record within a preset time period; determining a historical interaction context with the largest intersection exceeding a preset scenario adaptation threshold as a target historical interaction context, and taking a historical simplified information template type corresponding to the target historical interaction context as the target simplified information template.

[0017] By adopting the above technical solutions, the historical interaction context matching mechanism provides intelligent decision support for cases where the user is difficult to choose or has no clear preference. By calculating the intersection of the current set of information type labels and the historical interaction context, the system can identify similar historical experiences to the current situation and extract the user's preference patterns and successful experiences from them. The preset scenario adaptation threshold ensures the effectiveness of the historical reference, avoiding the interference of irrelevant historical data. This decision-making mechanism based on historical experience makes full use of the user's personal use habits and preference accumulation, realizing personalized intelligent recommendation. When the user cannot or does not want to make an active choice, the system can still make a reasonable simplified decision based on historical data, ensuring the continuity and consistency of the user experience, while avoiding interaction interruption or user experience degradation caused by selection difficulty.

[0018] In some embodiments in combination with the first aspect, before the step of obtaining an emotion feature event sequence based on the facial micro-expression data of the target user within a preset time window in a case where it is determined that the daily accumulation index is higher than the preset fatigue threshold, the method further comprises: taking the difference between the average value and the standard deviation of the daily accumulation index of the target user in the past preset statistical period as the preset fatigue threshold.

[0019] By adopting the technical solution, the dynamic calculation mechanism of the personalized fatigue threshold fully considers individual differences and habit changes of the user. By analyzing the daily accumulation index distribution of the user in the past preset statistical period, the average value and the standard deviation are calculated, and the difference between the average value and the standard deviation is taken as the fatigue threshold, so that the adaptive threshold setting based on the personal historical data is realized. The statistical method can identify the personal baseline level and variation range of the user, and avoid the misjudgment problem caused by the fixed threshold. For the user with regular screen use habits, the threshold setting is more accurate; for the user with large use mode change, the threshold has better adaptability. The personalized threshold ensures the accuracy of fatigue detection, reduces the occurrence of false positives and false negatives, provides reliable triggering conditions for subsequent emotion monitoring and content optimization, and improves the precision and user satisfaction of the entire adaptive system.

[0020] In a second aspect, the present application provides an AR glasses-based information adaptive presentation device, comprising: one or more processors and a memory; the memory is coupled to the one or more processors, and the memory is configured to store computer program code, the computer program code comprising computer instructions, and the one or more processors invoke the computer instructions to enable the AR glasses-based information adaptive presentation device to perform the method described in the first aspect and any possible implementation manner of the first aspect.

[0021] In a third aspect, the present application provides a computer program product comprising instructions, which, when executed on an AR glasses-based information adaptive presentation device, enable the AR glasses-based information adaptive presentation device to perform the method described in the first aspect and any possible implementation manner of the first aspect.

[0022] In a fourth aspect, the present application provides a computer-readable storage medium comprising instructions, which, when executed on an AR glasses-based information adaptive presentation device, enable the AR glasses-based information adaptive presentation device to perform the method described in the first aspect and any possible implementation manner of the first aspect. The one or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. By using the technical means of calculating the daily accumulation index based on cross-device usage data to evaluate the user's fatigue state, combining facial micro-expression data for emotion recognition, statistically analyzing negative emotional events and calculating the content acceptance index based on the negative degree weighting, and triggering information simplification display switching operation, when the user's screen usage intensity exceeds the personal historical average level, the system can actively monitor the user's emotional state changes, evaluate the user's acceptance of the current information content by quantitatively analyzing the frequency and duration of negative emotions, and automatically simplify the complexity and density of information display when detecting negative reactions, effectively solving the cognitive overload and visual fatigue problems caused by the fixed fixation time trigger mechanism of AR glasses in the prior art, and further realizing the technical effect of dynamically adjusting the information presentation strategy according to the user's real-time physiological and psychological state.

[0023] 2. The introduction of the sleep recovery factor can more accurately reflect the user's true fatigue state. The ratio of the actual sleep duration to the average sleep demand duration reflects the user's sleep adequacy, and the sleep quality score reflects the effectiveness of the sleep, and the sleep recovery factor obtained by multiplying the two comprehensively evaluates the user's physiological recovery ability. The ratio calculation of the daily accumulation index and the sleep recovery factor realizes the dynamic correction of the user's fatigue accumulation state, so that the fatigue evaluation not only considers the daily screen usage intensity, but also considers the user's recovery ability difference. This personalized fatigue evaluation mechanism can more accurately identify the user's true fatigue level, avoid underestimating the fatigue state due to sleep deficiency, or ignoring the good recovery state due to high-quality sleep, thereby improving the accuracy and effectiveness of subsequent emotion monitoring and content optimization decisions.

[0024] 3. The historical interaction context matching mechanism provides intelligent decision support for users who are difficult to choose or have no clear preferences. By calculating the intersection of the current information type tag set and the historical interaction context, the system can identify the most similar historical experience to the current situation and extract the user's preference pattern and successful experience from it. The preset scenario adaptation threshold ensures the effectiveness of the historical reference, avoiding interference from irrelevant historical data. This historical experience-based decision-making mechanism fully utilizes the user's personal usage habits and preference accumulation, realizing personalized intelligent recommendation. When the user cannot make an active choice, the system can still make a reasonable simplified decision based on historical data, ensuring the continuity and consistency of the user experience, while avoiding interaction interruption or user experience degradation caused by choice difficulty. BRIEF DESCRIPTION OF DRAWINGS

[0025] Figure 1 is a flowchart of an information adaptive presentation method based on AR glasses in an embodiment of the present application; Figure 2is another flowchart of a method for adaptive information presentation based on AR glasses in an embodiment of the present application; Figure 3 is an exemplary hardware structure diagram of a device for adaptive information presentation based on AR glasses in an embodiment of the present application. DETAILED DESCRIPTION

[0026] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments and are not intended to be limiting of the present application. As used in the specification and the appended claims of the present application, the singular forms "a," "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "and / or," as used in the present application, signify and include any and all possible combinations of one or more of the associated listed items.

[0027] Hereinafter, the terms "first" and "second" are only for the purpose of description and cannot be understood as implying or suggesting relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined with "first" and "second" can explicitly or implicitly include one or more of the features, and in the description of the embodiments of the present application, the meaning of "a plurality of" is two or more, unless otherwise specified.

[0028] For ease of understanding, the related terms and concepts involved in the embodiments of the present application are introduced first as follows.

[0029] AR glasses (Augmented Reality Glasses) are a kind of wearable augmented reality device that enhances the perception of the real world by superimposing virtual information in the user's field of view. AR glasses usually contain display screens, cameras, sensors, processors and other core components, which can capture the user's visual environment in real time and superimpose digital content in it. Unlike traditional VR devices, AR glasses allow users to see both the real world and virtual information, realizing a virtual-real integrated interactive experience. In the present application, AR glasses have functions such as facial micro-expression recognition, eye tracking, cross-device data synchronization, etc.

[0030] Please refer to Figure 1 is a flowchart of a method for adaptive information presentation based on AR glasses in an embodiment of the present application.

[0031] S101、At the beginning of the target user's current session, the ratio of the target user's total screen exposure time on the day to the average total screen exposure time in the preset historical period is calculated based on the acquired cross-device usage data, and the ratio is taken as the daily degree accumulation index.

[0032] Wherein, the target user represents a specific individual wearing and using the AR glasses; the current session refers to the complete time period from the user's current start and beginning of interaction with the AR glasses until the end of the interaction or the device standby; the cross-device usage data refers to the usage time data recorded by other electronic devices (such as smartphones, tablets, personal computers) bound to the target user's identity account; the total screen exposure time is used to represent the total time of the target user's screen on in all associated electronic devices in the day, which is the cumulative value of the target AR glasses' daily usage time and the total usage time of the associated terminal devices associated with the target user in the day, for example, if the user has used the mobile phone for 2 hours and the AR glasses for 0.5 hours in the day, the total screen exposure time is 2.5 hours; the preset historical period refers to a specific time span for calculating the historical average, such as the past 30 days or 90 days; the average total screen exposure time refers to the arithmetic mean of the user's daily total screen exposure time in the preset historical period.

[0033] Specifically, in the initial stage of the target user wearing and activating the AR glasses to start a new interaction session, the device will perform this step to evaluate the user's initial fatigue baseline. The device first obtains the cumulative usage time of itself from zero o'clock of the day to the current time through the logs or timers recorded by its operating system. At the same time, the device accesses a cloud service platform associated with the user's account through network connection (such as Wi-Fi or Bluetooth), or directly communicates with the user's other associated terminal devices (for example, smartphones with installed specific synchronization applications) through local network protocols, to obtain the total screen usage time recorded by these devices in the day (for example, by calling the digital health API interface or screen time statistics service interface of the operating system). The device sums up the daily usage time of all devices to obtain the total screen exposure time in the day. Then, the device retrieves the total screen exposure time data of each day in the preset historical period (for example, the past 30 days) from the local storage or cloud database, and calculates the arithmetic mean of these data to obtain the average total screen exposure time. Finally, the device divides the total screen exposure time in the day by the average total screen exposure time to obtain a dimensionless ratio, and takes this ratio as the initial daily accumulation index. This index intuitively reflects whether the user's screen usage in the day exceeds his personal regular level.

[0034] In some embodiments, after calculating the initial daily accumulation index, the device further obtains physiological data related to the user's health for correction. The device obtains the actual sleep duration of the target user the previous night, the average sleep demand duration calculated from historical data, and the sleep quality score generated by the health application based on the associated terminal device (such as a smart bracelet or smart watch). Then, the device calculates the ratio of the actual sleep duration to the average sleep demand duration, and multiplies the ratio by the sleep quality score (usually a normalized value between 0 and 1) to obtain a comprehensive sleep recovery factor. Finally, the device divides the initial daily accumulation index by the sleep recovery factor to obtain an updated daily accumulation index, which more accurately reflects the user's true physiological fatigue accumulation state considering the recovery effect of the night's rest. Optionally, the device can also introduce usage scenario weights for adjustment. The device identifies and classifies the user's usage behavior on each device (e.g., high-intensity work programming on a PC, leisurely social media browsing on a mobile phone), and assigns different fatigue weight coefficients to different scenarios (e.g., 1.5 for work programming and 0.8 for leisure browsing). When calculating the total screen exposure duration, the device multiplies the usage duration of each device by its corresponding scenario weight coefficient and then accumulates them to obtain a weighted total screen exposure duration, which is used as the basis for calculating the daily accumulation index, so that the index can more accurately reflect the cognitive load differences brought by different activities. It can be understood that other ways can be used to calculate the daily accumulation index, such as introducing user subjective feedback for calibration, which is not limited here.

[0035] S102, in the case where the daily accumulation index is higher than the preset fatigue threshold, obtaining an emotion feature event sequence based on the facial micro-expression data of the target user within a preset time window and combining an emotion recognition algorithm.

[0036] wherein the facial micro-expression data is captured by the camera in the AR glasses that faces the user's face; the emotion recognition algorithm usually refers to a pre-trained deep learning model (such as a convolutional neural network CNN) that can map the input micro-expression data to a specific emotion category; the emotion feature event sequence is an ordered list with timestamps that records all emotion events and their attributes (such as emotion type, start time, duration, intensity) recognized within the time window, for example [(t1, 'confusion', 0.8s), (t2, 'frustration', 1.2s)].

[0037] After calculating the daily fatigue accumulation index in S101, the device compares it with a preset fatigue threshold. Only when the index is higher than the threshold, the device activates the micro-expression monitoring module in this step. The module uses the camera facing the user's face, the device activates the built-in front camera module, and starts to continuously collect high-resolution image data of the user's face area, and the collection frequency is usually set to 15 to 30 frames per second to ensure that subtle expression changes can be captured. The device divides these image data into segments of the same length as the preset time window for processing. In each time window, the device inputs the collected facial image data into a pre-trained emotion recognition algorithm model, which can recognize multiple emotional states including joy, disgust, fatigue, anxiety, etc., and assign a corresponding confidence score to each emotion. The algorithm model outputs the emotion recognition result in real time by analyzing the position change of the facial key points, the eye blinking frequency, the degree of eyebrow wrinkling, the direction of mouth bending, etc. The device organizes the continuous emotion recognition results into an emotion feature event sequence in chronological order, and each event contains detailed information such as emotion type, start time, end time, and emotion intensity.

[0038] In some embodiments, the setting of the fatigue threshold can be achieved in various ways: optionally, the device dynamically calculates and updates the preset fatigue threshold based on the daily fatigue accumulation index of the target user in a preset statistical period (e.g., the last 90 days). Specifically, the device calculates the arithmetic mean and standard deviation of all daily fatigue accumulation indexes in this period. Then, the device takes the difference between the mean and the standard deviation (i.e., mean - standard deviation) as the personalized preset fatigue threshold for this user. This adaptive threshold setting method can take into account the differences between users and the long-term changes in individual habits, making the fatigue judgment more accurate and personalized. Optionally, the device sets the threshold by establishing a calibration mechanism based on user subjective feedback. Specifically, the device asks the user for their current subjective fatigue level (e.g., 'not tired','slightly tired', 'tired') in a non-disturbing way (such as a simple UI prompt) after a long period of use or at the end of a conversation. The device stores the user's feedback fatigue level in association with the corresponding daily fatigue accumulation index at that time. When enough data points are collected, the device trains a classifier through a machine learning model (such as a support vector machine or logistic regression), and the decision boundary of the classifier is used as the personalized preset fatigue threshold, so that the threshold can reflect the individual physiological perception of the user, which is not limited here.

[0039] In some embodiments, the generation of the sequence of emotional feature events can be achieved in several ways. Optionally, the device captures a high frame rate video stream through the facial camera and locates the face bounding box in each frame using a face detection algorithm (e.g., the multi-task cascaded convolutional neural network, MTCNN). Secondly, within the face region, the device applies a facial landmark detection model to track the 2D coordinates of dozens of landmarks (e.g., eyebrow contour points, eye corner points, nose tip, lip contour points) in real time. Thirdly, the device calculates the displacement vectors of the landmark coordinates between consecutive frames and extracts dynamic features such as the average displacement velocity, acceleration, and relative position change rate of a specific group of landmarks (e.g., the region corresponding to the "frown muscle") to form a feature vector. Finally, these dynamic feature vectors are input into a pre-trained time series neural network (e.g., the long short-term memory network, LSTM) that is trained to recognize specific micro-expression patterns related to cognitive load and output emotional events with confidence, forming a sequence. Optionally, the device uses an analysis method based on facial action coding units (AUs) to generate the sequence. Specifically, the device first detects the activation state and intensity of multiple key AUs in the user's face in real time through a pre-trained model, such as AU4 (frown), AU7 (eyelid tightness), AU12 (mouth corner up), etc. Then, the device inputs these time-varying AU activation sequences into a rule engine or a shallow classification model. The rule engine maps specific AU combination patterns to emotional events (e.g., detects high-intensity activation of AU4 and AU7 at the same time, then generates an 'irritated' or 'difficulty concentrating' emotional event) according to the theory of the facial action coding system (FACS), and records their start and end times and intensities, ultimately forming a sequence of emotional feature events.

[0040] It should be noted that, in order to protect user privacy, all facial image data is processed locally on the device and is not uploaded to a cloud server, and the emotional recognition algorithm model is processed using differential privacy technology to protect the user's specific facial feature information.

[0041] It can be understood that the emotional recognition algorithm and the face detection algorithm are prior art and will not be described here.

[0042] S103, count the number of negative emotional feature events and the cumulative duration of the emotional feature events, and calculate the content acceptance index based on the negative degree of the emotional feature events.

[0043] Among them, the negative emotional feature event refers to the event record identified as expressing a negative emotional state in the sequence of emotional feature events, including emotions such as disgust, fatigue, anxiety, and anger.

[0044] After the device obtains the sequence of mood feature events generated in S102, the sequence will be parsed and quantified. The device will iterate through each mood feature event in the sequence and categorize it as "negative", "neutral", or "positive" according to the event's type label. The device will filter out all events that are labeled as "negative". Then, the device will perform two basic statistics on these negative events: one is to count their total number, resulting in the occurrence frequency; the other is to add up the duration of each negative event (end timestamp - start timestamp), resulting in the cumulative duration. Next, the device will look up the corresponding negative degree weight value for each specific negative mood type (such as 'irritated', 'confused') from a pre-set weight configuration library. The device will use a weighted calculation to integrate these data, for example, the content acceptance index = 1 / (1 + a * Σ (occurrence frequency_i * weight_i) + β * Σ (duration_j * weight_j)), where i and j iterate through all negative mood types, and a and β are adjustment coefficients. It can be understood that the weight configuration library associates the algorithm-recognizable negative mood label with a quantified numerical value representing the degree of negative impact on user experience, and these weight values are usually set based on psychological research or large-scale user experiment data.

[0045] In some embodiments, the calculation of the content acceptance index can be implemented in various ways. Optionally, the device aligns the sequence of emotional feature events generated in S102 with the gaze point data synchronously acquired by the eye tracking module based on the common timestamps. For each negative emotional feature event, the device determines which specific information content unit (e.g. a text paragraph, an image, or a data chart) on the screen the user’s gaze point falls on during the event, thereby binding the negative emotion to the specific content. Then, the device calculates the negative feedback rate for each information content unit that is “bound” with negative emotion, which is equal to the cumulative duration of all negative emotional events associated with the unit divided by the total duration of the user’s gaze on the unit. Finally, the device weights the negative feedback rates of all information content units based on the information weight of each unit in the current information panel (e.g. determined according to its display area or text length), and sums them up to obtain the final content acceptance index. This way can reflect which specific content causes the user’s negative reaction. Optionally, the device considers the dynamic evolution trend of the emotion. Instead of only counting the stock of negative emotion, the device also analyzes its increment, i.e. the rate of emotional change. The device calculates the derivative or difference of the negative emotional intensity (e.g. the confidence output by the algorithm) within a time window to identify the “burst point” of the emotion. The calculation formula of the content acceptance index introduces an additional penalty term that is positively correlated with the maximum instantaneous growth rate of the negative emotional intensity. In this way, a sudden and intense burst of negative emotion will lower the content acceptance index more significantly than a long-term and mild negative emotion, making the system more sensitive to the user’s sudden discomfort reaction. It can be understood that other ways can also be used to calculate the content acceptance index, such as introducing historical interaction data to give higher weight to the penalty of the content pattern that repeatedly causes negative emotions for a specific user, which is not limited here.

[0046] S104、In response to determining that the content acceptance index is lower than the preset negative reaction threshold, triggering an information simplification display switching operation.

[0047] The information simplification display switching operation refers to the device actively changing the presentation of the currently displayed information content, reducing the quantity, complexity or abstraction level of the information to reduce the user’s cognitive load. For example, switching from a complete product information panel containing detailed specifications, user reviews and picture-text introductions to a summary view showing only the core selling points and price.

[0048] After calculating the real-time content acceptance index at S103, the device compares it with a preset negative reaction threshold. The threshold can be a fixed empirical value or a personalized value dynamically adjusted according to historical feedback of the user. It can be understood that if the device determines that the content acceptance index is lower than the threshold, it means that the current information presentation method has caused cognitive stress or negative emotions to the user, and the device will trigger the information simplification display switching operation. Specifically, the device first analyzes the structural composition and content hierarchy of the information panel currently being displayed, identifies various information components such as text information, image elements, three-dimensional models, interactive buttons, etc. contained therein. The device classifies and sorts the currently displayed information content according to the preset information importance priority rules, marks the core key information as high priority, marks the auxiliary explanatory information as medium priority, and marks the decorative or redundant information as low priority. The device performs the information simplification display operation, generates a simplified version of the information display interface by hiding low-priority information, compressing the display space of medium-priority information, and simplifying the expression of high-priority information. The simplified interface retains the core information content, reduces the number and complexity of visual elements, reduces the cognitive processing burden of the user, while maintaining the integrity and usability of the information, helping the user to continue information acquisition and interaction operation under lower cognitive stress.

[0049] In some embodiments, the trigger and execution of the information simplification display switching operation can be implemented in various ways. Optionally, the device uses natural language processing technology to perform semantic analysis and keyword extraction on the currently displayed text content, identifies the subject category and importance of the information, and then dynamically adjusts the information priority sorting rules according to the user's historical browsing preferences and current task goals, and then uses a gradual simplification method to hide unimportant decorative elements first, then compress the display space of secondary information, and finally simplify the expression form of core information. The whole process is presented to the user through smooth animation transition effect, avoiding user confusion caused by abrupt interface changes. Optionally, the device implements a precise simplification mechanism based on user gaze heat map, monitors the user's gaze focus distribution in real time through eye tracking technology, generates a gaze heat map of the current information panel, identifies information areas with high user attention and completely ignored information areas, and then preferentially retains the information content of high-attention areas and gradually removes or folds the information elements of low-attention areas. At the same time, according to the user's gaze path and dwell time, dynamically adjust the information layout and display order to ensure that the simplified interface is more in line with the actual needs and browsing habits of the user. It can be understood that other ways of implementing simplification switching can also be used, such as providing a scrollable "information density" slider for the user to manually adjust, etc., which are not limited here.

[0050] It should be noted that the information simplification display switching operation also needs to consider the context association and logical integrity of the information. The device analyzes the dependency and reference relationship between the information elements to ensure that the logical structure and semantic coherence of the information are not damaged in the simplification process, while providing a quick recovery mechanism to allow the user to expand the hidden information content when needed.

[0051] In the embodiments of the present application, since the technical solutions of fusing cross-device usage data to calculate daily accumulation index, based on facial micro-expression data to identify emotions, statistical negative emotional feature events to calculate content acceptance index, and triggering information simplification display switching operation are adopted, when the user's screen usage intensity exceeds the personal historical average level, the device can actively identify the emotional state changes of the user, evaluate the user's acceptance degree of the current information content by quantitatively analyzing the occurrence frequency and duration of negative emotions, and automatically simplify the complexity and density of information display when detecting negative reactions of the user, effectively solving the cognitive overload and visual fatigue problems caused by the fixed fixation time trigger mechanism of AR glasses in the prior art, and further realizing the technical effect of dynamically adjusting the information presentation strategy according to the real-time state of the user, and improving the information acquisition efficiency of the user when using the AR device for a long time.

[0052] In the above embodiments, the device can realize the technical effect of dynamically adjusting the information presentation strategy according to the real-time physiological and psychological state of the user by fusing cross-device usage data to calculate daily accumulation index, based on facial micro-expression data to identify emotions, statistical negative emotional feature events to calculate content acceptance index, and triggering information simplification display switching operation. In actual application, when the above method is executed, there may be a defect that it is not possible to accurately identify which specific information content triggers the negative emotional reaction of the user, resulting in lack of pertinence of the simplification operation, affecting the accuracy of information optimization and the improvement effect of user experience.

[0053] Please refer to Figure 2 is another flowchart of an information adaptive presentation method based on AR glasses in the embodiments of the present application.

[0054] S201, at the beginning of the current session of the target user, calculating the ratio of the total screen exposure time of the target user on the current day to the average total screen exposure time in the preset historical period based on the obtained cross-device usage data, and taking the ratio as a daily accumulation index.

[0055] S202, in the case where the daily accumulation index is higher than the preset fatigue threshold, obtaining an emotional feature event sequence based on the facial micro-expression data of the target user in the preset time window combined with an emotion recognition algorithm.

[0056] Steps S201-S202 are combined with Figure 1The steps S101-S102 in the illustrated embodiment are similar, and will not be described again here.

[0057] S203, time-stamp align the sequence of emotional feature events with the synchronous gaze point data acquired through eye tracking, and bind one or more information content units that the target user gazes at at the corresponding time for each negative emotional feature event.

[0058] Wherein, the eye tracking refers to a technology for monitoring the eye movement trajectory and gaze position of the user in real time through an infrared camera or other sensors, the synchronous gaze point data is used to represent the user's visual focus coordinate information collected synchronously in time with the sequence of emotional feature events, and the information content unit represents an independently identifiable information display block in the current AR interface, such as a single text paragraph, a picture, a button, or a data table, etc.

[0059] When the device obtains the sequence of emotional feature events, the device first acquires the gaze point coordinate sequence completely synchronized with the emotional data collection time period from the eye tracking module, which records the real-time position change of the user's visual line on the screen. The device uses a time-stamp matching algorithm to align the time marker of each event in the sequence of emotional feature events with the time marker of the gaze point data, and handles possible time deviation and sampling frequency difference. The device converts the two-dimensional coordinates of the gaze point into corresponding information content unit identifiers through screen coordinate mapping technology, and identifies the specific information area that the user is watching at a specific time. For each negative emotional feature event, the device will find the user's gaze trajectory in the time period when the event occurs, determine one or more information content units that the user mainly focuses on in that time period, and establish a binding relationship record between the negative emotional event and these information content units in the database.

[0060] In some embodiments, the binding of emotional events and gaze content can be achieved in various ways. Optionally, the device uses a sliding window time alignment algorithm, first sets a time tolerance range (such as ±100 milliseconds), for each negative emotional feature event, the device searches for the closest gaze point data within the tolerance range before and after its start time, then uses a linear interpolation method to estimate the gaze position at the emotional event time, then determines which specific information content unit the gaze point falls into through the bounding box detection algorithm of the interface element, finally establishes a many-to-many mapping relationship between the emotional event ID and the information content unit ID, and records the binding confidence score. Optionally, the device generates an attention heat map for the time period by analyzing the complete gaze trajectory of the user during the negative emotional event, identifies the information area where the user's line of sight stays the longest and visits the most frequently, and then assigns weight coefficients to different information content units according to gaze duration and visit frequency, and distributes the negative emotional event to multiple related information content units in proportion to the weight, thereby establishing the emotion-content association. It can be understood that other ways of binding emotional events and content can also be used, which are not limited here.

[0061] S204, calculate the ratio of the cumulative duration of the negative emotional feature events associated with the information content unit to the total duration of the target user's gaze, to obtain the negative feedback rate of the information content unit.

[0062] After the binding of emotional events and information content units is completed, the device needs to quantitatively evaluate the specific impact of each information content unit on the user's emotions. Specifically, the device first traverses all the records of established binding relationships, for each information content unit, it counts the duration of all negative emotional feature events associated with it, and sums up these durations to obtain the negative emotional cumulative duration of the information content unit. At the same time, the device extracts the complete gaze record of the user on the information content unit from the eye tracking data, calculates the sum of all time periods when the user's line of sight falls within the unit area, and obtains the total gaze duration of the user on the unit. The device divides the negative emotional cumulative duration by the total gaze duration to obtain a ratio between 0 and 1, which is the negative feedback rate of the information content unit. The higher the ratio, the greater the proportion of negative emotions generated by the user when watching the content, and the more significant the negative impact of the content on the user. The device records and sorts the negative feedback rates of all information content units to provide quantitative basis for subsequent content optimization and simplification operations.

[0063] In some embodiments, the calculation of the negative feedback rate can be implemented in various ways. Optionally, the device adopts a time window segmentation method to segment the user's gaze duration according to a fixed time interval (e.g., 5 seconds), calculates the negative emotional intensity in each time segment, then calculates the ratio of the weighted negative emotional duration to the gaze duration in the time segment, and finally calculates the weighted average of the ratios of all time segments to obtain a more accurate negative feedback rate. This method can better reflect the influence of emotional intensity changes on the feedback rate. Optionally, the device introduces a gaze quality evaluation mechanism to evaluate the effectiveness of the gaze by analyzing the user's eye movement patterns (such as scanning video rate, gaze stability, pupil changes, etc.), and eliminates or reduces the weight of the low-quality gaze duration (such as gaze during rapid scanning or distraction) from the total gaze duration, then recalculates the negative feedback rate based on the corrected effective gaze duration, thereby improving the accuracy and reliability of the calculation results. It can be understood that other ways of calculating the negative feedback rate can also be used, which are not limited here.

[0064] S205, based on the information amount weight of the information content unit in the current information panel, the negative feedback rate is weighted and summed to obtain the content acceptance index.

[0065] Among them, the information amount weight represents the quantitative value of the importance or influence size of each information content unit in the current information panel, which is usually determined based on factors such as display area, text length, position importance or semantic importance of the content.

[0066] When the device obtains the negative feedback rates of all information content units, it needs to integrate these scattered evaluation results into a comprehensive index reflecting the overall content acceptance. Specifically, the device first analyzes the layout structure and content composition of the current information panel, identifies all information content units contained therein, and assigns each unit a corresponding information amount weight according to the preset weight calculation rule. The weight allocation usually considers multiple dimensions, including the display area proportion of the content unit on the screen, the number of characters in the text content, the size of the image pixels, the position importance in the page layout (such as the title area weight is higher than the note area), and the semantic importance of the content. The device multiplies the negative feedback rate of each information content unit by its corresponding information amount weight to obtain the weighted negative feedback value of the unit. The device adds up all the weighted negative feedback values of the information content units, then normalizes them by dividing by the sum of all information amount weights to obtain a content acceptance index between 0 and 1. The closer the index is to 0, the higher the user's acceptance of the current content, and the closer the index is to 1, the stronger the user's negative reaction.

[0067] In some embodiments, the allocation of information weight and the calculation of content acceptance index can be implemented in various ways. Optionally, the device employs a multi-dimensional weight fusion algorithm, first calculates the visual weight of each information content unit based on visual saliency theory, determines the visual appeal of the content by analyzing visual features such as color contrast, brightness distribution, and edge density, then calculates the information entropy weight of the content based on information theory, determines the information value of the content by analyzing the word frequency distribution, semantic complexity, and information density of the text, then calculates the personalized weight combining the user's historical interaction data, adjusts according to the user's attention preference and interaction frequency of different types of content, and finally fuses the weights of multiple dimensions through weighted average or product operation to obtain the comprehensive information weight. Optionally, the device implements a dynamic weight adjustment mechanism, adjusts the weight allocation strategy in real time according to the current usage context and task target, for example, increases the weight of price and evaluation information in the shopping scene, increases the weight of core concepts and key knowledge points in the learning scene, and at the same time uses machine learning algorithm to analyze the user's real-time feedback and behavior pattern, automatically optimizes the weight allocation parameters, so that the content acceptance index can more accurately reflect the user's real feelings and needs in a specific situation. It can be understood that other ways of weight allocation and index calculation can also be used, which are not limited here. It should be noted that when there are dynamic content or interactive elements in the information panel, the device will adjust the weight accordingly according to the update frequency and interaction complexity of the content.

[0068] S206、In the case where the content acceptance index is determined to be lower than the negative reaction threshold, a pre-trained natural language processing model is called to perform semantic analysis on the currently displayed text content to obtain a semantic analysis result.

[0069] wherein the pre-trained natural language processing model refers to a deep learning model trained on large-scale text data, with text understanding and semantic analysis capabilities, semantic analysis refers to a deep linguistic analysis of text content, extracting its semantic structure, theme information, and key concepts, semantic analysis result represents the structured semantic information output by the model, including text theme classification, key entities, semantic relationships, and other multi-dimensional analysis results, information type label is used to represent the functional category or theme category to which the text content belongs, and key semantic entity refers to a word or phrase with important semantic value in the text, such as product name, technical parameter, user evaluation, etc.

[0070] When the device detects that the content acceptance index is below the preset negative reaction threshold, it indicates that the current information display has caused cognitive burden to the user, and the intelligent content optimization process needs to be started. Specifically, the device first extracts all the text content in the current information panel, including the title, body, label, button text, and various text elements, and integrates these text contents into a text dataset for analysis. The device calls a pre-trained natural language processing model built-in or deployed in the cloud, which is usually based on the Transformer architecture (such as BERT, GPT, etc.), and has powerful text understanding and semantic analysis capabilities. The model performs multi-level semantic analysis on the input text content, first performing lexical analysis and syntactic analysis to identify the basic language structure of the text, then performing semantic role labeling and entity recognition to extract key semantic entities in the text and their relationships, then performing topic classification and sentiment analysis to determine the main theme and emotional tendency of the text, and finally outputting structured semantic analysis results, including information type labels (such as product description, technical specifications, user feedback, etc.) and a list of key semantic entities.

[0071] In some embodiments, semantic analysis and result extraction of text content can be achieved in various ways. Optionally, the device uses a multi-model integrated semantic analysis scheme, first uses a named entity recognition model to extract specific entities such as names, places, product names, and times in the text, then uses a text classification model to perform multi-label classification on the text content to identify its information type label, then uses a relation extraction model to analyze the semantic relationships between entities to construct a knowledge graph structure, and finally uses a sentiment analysis model to evaluate the sentiment polarity and intensity of the text, and fuses and processes the output results of all models to generate comprehensive semantic analysis results. Optionally, the device implements a domain adaptation-based semantic analysis strategy, selects a corresponding domain-specific language model according to the current application scenario (such as e-commerce shopping, education and learning, entertainment and leisure, etc.), which is fine-tuned and optimized for specific domain terminology and expression based on a general language model, can more accurately identify domain-related information type labels and key semantic entities, and uses an incremental learning mechanism to continuously optimize the analysis accuracy and adaptability of the model based on user interaction feedback. It can be understood that other ways of implementing semantic analysis can also be used, which are not limited here.

[0072] It should be noted that, in order to protect user privacy and improve processing efficiency, the device preferentially uses a locally deployed lightweight language model for semantic analysis, and only calls a cloud-based large-scale model for complex text processing or high-precision analysis.

[0073] S207, after obtaining the negative feedback rate of all information content units, determine the highest negative feedback rate among all information type labels as the high negative feedback information category and generate a first simplified information template.

[0074] After obtaining the negative feedback rate data of all information content units, the device needs to identify the main information categories that cause negative reactions of the user and generate a corresponding content optimization scheme. Specifically, the device first groups all information content units according to the information type tags identified in S206, for example, grouping all content units marked as "product description category" into one group, and grouping all content units of "technical specification category" into another group. The device calculates the average negative feedback rate of each information category, that is, the arithmetic mean or weighted mean of the negative feedback rates of all information content units in the category. The device sorts the average negative feedback rates of all information categories and identifies the information category with the highest negative feedback rate as the high-negative-feedback information category. The device generates a first simplified information template, which simplifies information by completely deleting or hiding all content units belonging to the high-negative-feedback information category from the current information panel, and retains information content of other categories, thereby reducing the user's exposure to information elements that trigger negative emotions and reducing overall cognitive burden and emotional stress.

[0075] In some embodiments, the identification of the high-negative-feedback information category and the generation of the first simplified information template can be achieved in various ways. Optionally, the device uses a hierarchical information category analysis method, first identifies the main information types (such as descriptive information, specification information, and evaluation information) at a coarse-grained level, and then further subdivides the subcategories of each main type (such as product description, which can be subdivided into function description, appearance description, and use scenario description) at a fine-grained level, calculates the negative feedback rate of each level, uses a multi-level decision mechanism to determine the information category with the most negative impact, and generates simplified template options of different granularities according to the hierarchical relationship of the categories. Optionally, the device implements a dynamic threshold adjustment mechanism to dynamically adjust the negative feedback rate threshold according to the user's personal preferences and historical interaction data, and for users who are particularly sensitive to a certain type of information, the threshold for marking this type of information as high-negative-feedback is lowered. At the same time, considering the importance weight of the information category, avoid deleting information content that is crucial to task completion, and use a gradual deletion strategy when generating the first simplified information template, preferentially deleting information categories with lower importance and higher negative feedback rate. It can be understood that other ways of implementing information category analysis and template generation can also be used, which are not limited here. It should be noted that when deleting the high-negative-feedback information category, the device will retain the key information summary or provide a quick access entry in this category to ensure that the user can still access the complete information when needed.

[0076] S208, based on the negative feedback rate, determining all information categories with a negative feedback rate lower than a preset safety threshold as a set of low-negative-feedback information categories, and generating a second simplified information template.

[0077] After identifying high-negative-feedback information categories, the device needs to generate a simplified solution from another perspective: retaining and optimizing information content with high user acceptance. Specifically, the device iterates through the negative feedback rate data of all information categories, filtering out information categories with negative feedback rates below a preset safety threshold to form a low-negative-feedback information category set. The preset safety threshold is usually dynamically determined based on the user's personal characteristics, historical data, and current status; for example, it can be set to 0.5 times the user's historical average negative feedback rate or a fixed value such as 0.3. The device reorganizes and optimizes the layout of all information content units in the low-negative-feedback information category set, using an information aggregation algorithm to merge and display highly relevant content, and a priority sorting algorithm to place the most important and most popular content in prominent positions. The device generates a second simplified information template, which simplifies information by retaining information categories that users have positive feedback on and optimizing their display, while ensuring the integrity and accessibility of key information, providing users with a browsing experience that is less cognitively burdensome and has higher information value.

[0078] In some embodiments, the filtering of low-negative feedback information categories and the generation of a second simplified information template can be achieved through various methods. Optionally, the device employs an adaptive threshold determination mechanism. First, it analyzes the distribution of users' historical negative feedback rates, calculates the mean and standard deviation, and then subtracts one standard deviation from the mean to obtain a personalized preset safety threshold. For users with high emotional sensitivity, the threshold is further lowered to ensure that the filtered low-negative feedback information categories truly match the user's acceptance level. When generating the second simplified information template, an intelligent layout algorithm is used to optimize the arrangement order and visual hierarchy of information based on the semantic relevance of the information content and the user's reading habits. Optionally, the device implements a content quality assessment and optimization mechanism. It further analyzes the content in the low-negative feedback information categories, identifies content with room for optimization through methods such as text readability assessment, information density analysis, and visual complexity calculation, and then refines and optimizes this content using automatic summarization technology, keyword extraction algorithms, and visual simplification technology to generate a high-quality information display scheme that maintains information integrity while reducing cognitive burden. It is understood that other methods can also be used to achieve information filtering and template optimization, which are not limited here. It should be noted that the second simplified information template retains a quick access path to the complete information, allowing users to view the simplified or hidden details when needed.

[0079] S209. Use the first simplified information template and the second simplified information template as candidate simplified information templates.

[0080] After generating the simplified information templates of the two different strategies, the device needs to integrate them into a candidate scheme set to provide a basis for subsequent user selection and system decision. Specifically, the device will standardize the first simplified information template generated in S207 and the second simplified information template generated in S208, ensuring that the two templates are consistent in format, layout and interaction mode, facilitating user comparison and selection. The device assigns a unique identifier and description label to each template, for example, marking the first simplified information template as "simplified version" and the second simplified information template as "preferred version". The device records and stores the metadata information of the two templates (such as the type of information contained, the expected cognitive burden level, the information completeness, etc.) to provide data support for subsequent intelligent recommendation and personalized optimization. The device establishes a management mechanism for candidate simplified information templates, including template preview generation, quality evaluation and user adaptability analysis, etc. to ensure that the candidate templates can effectively meet the different needs and preferences of users.

[0081] S210, based on the target user's gaze duration on the candidate simplified information template within the preset response time, determining the simplified information template whose gaze duration exceeds the preset selection threshold as the target simplified information template.

[0082] When the device generates candidate simplified information templates, it needs to determine the most suitable scheme through the natural interaction behavior of the user, rather than forcing the user to make an explicit selection. Specifically, the device displays preview versions of the first simplified information template and the second simplified information template on the screen at the same time, usually in a side-by-side display or partition display manner, ensuring that the user can observe and compare the two options simultaneously. The device starts the eye tracking module to monitor the user's eye movement trajectory and gaze behavior within the preset response time, and records the user's gaze duration data for each candidate template in real time. The device sets a preset response time window (e.g. 5-10 seconds), within which it continuously collects the user's gaze data and calculates the cumulative gaze duration of the user for each candidate template in real time. When the preset response time ends, the device compares the gaze duration of each candidate template with the preset selection threshold, and determines the template whose gaze duration exceeds the threshold as the user's preferred choice. If the gaze duration of multiple templates exceeds the threshold, the template with the longest gaze duration is selected as the target simplified information template; if none of the templates has a gaze duration exceeding the threshold, a backup decision mechanism is triggered.

[0083] S211, performing information simplification display switching operation based on the target simplified information template.

[0084] After determining the target simplified information template, the device needs to perform actual interface switching and content reorganization operations to provide the user with an optimized information browsing experience. Specifically, the device first analyzes the specific content composition and layout requirements of the target simplified information template, determines the information elements that need to be retained, deleted, modified, or rearranged. The device performs content filtering operations, removes or hides unnecessary information content units from the current display interface according to the definition of the template, while retaining the information content specified in the template. The device performs layout reorganization, rearranges the display position, size, and visual hierarchy of the retained information according to the design requirements of the target template, optimizes the arrangement order and spatial distribution of the information. The device applies visual optimization processing, including adjusting font size, color contrast, spacing settings, and other visual parameters to ensure that the simplified interface has good readability and aesthetics. The device performs smooth animation transition effects, realizes natural switching from the original interface to the simplified interface through animation techniques such as gradient, sliding, or fade-in and fade-out, avoiding abrupt changes that disturb the user. The device updates the interaction logic of the interface, adjusts button functions, navigation paths, and operation processes to ensure that the user can normally complete the required interaction tasks in the simplified interface.

[0085] In some embodiments, the execution of the information simplification display switching operation can be implemented in various ways. Optionally, the device adopts a hierarchical and progressive switching strategy, decomposes the simplification switching process into multiple stages, first performs coarse-grained content filtering to remove obviously unnecessary information modules, then performs fine-grained content optimization to adjust the expression method and display density of the retained information, and finally performs visual beautification and interaction optimization, each stage is presented to the user through smooth animation effects, allowing the user to clearly perceive the interface change process and final effect. Optionally, the device implements an intelligent content replacement mechanism for high negative feedback information content that is deleted, instead of simply removing it, but replacing it with more concise and easier-to-understand alternative content, such as replacing complex text descriptions with icons, replacing detailed technical specifications with summaries, replacing lengthy user reviews with ratings, while providing an entry for quick access to the original detailed information to ensure information availability. It can be understood that other ways of implementing the simplification switching operation can also be used, which are not limited here.

[0086] S212、In the case where it is determined that the target user does not respond within the preset response time, the intersection of the information type tags in the historical interaction context and the current information type tag set is calculated in the historical interaction record based on the information type tag set contained in the currently displayed information within a preset time period.

[0087] When the user does not show a clear selection preference for the candidate simplified information templates within the preset response time, the device needs to start a backup decision mechanism to infer the most suitable simplification scheme by analyzing the user's historical behavior patterns. Specifically, the device first confirms that the user indeed does not generate effective selection behavior within the preset response time, that is, the gaze duration of all candidate templates does not exceed the preset selection threshold, or the user's gaze behavior is too scattered to form a clear preference judgment. The device extracts all information type tags contained in the current display information to form a current information type tag set, such as {product description class, technical specification class, user feedback class, price information class}. The device accesses the user's historical interaction record database, searches the user's historical interaction context records within a preset time period (such as the past 30 days or 90 days), and each historical context contains the information type tag set at that time, the user's selection behavior, emotional response, and final satisfaction evaluation, etc. The device calculates the intersection of the information type tag set in each historical interaction context and the current information type tag set one by one, counts the number of tags contained in the intersection, and identifies the historical context with the largest intersection and the intersection size exceeding the preset scenario adaptation threshold. The device uses these high-similarity historical contexts as a reference basis to analyze the user's preference patterns and selection behavior in similar situations, providing data support for the current simplification decision.

[0088] In some embodiments, the matching and analysis of historical interaction contexts can be implemented in various ways. Optionally, the device uses a weighted similarity calculation method, which not only considers the number of intersection of information type tags, but also assigns weights to each tag according to the importance of different tags and the user's personal preferences, calculates the weighted intersection similarity, and introduces a time decay factor to give higher weights to more recent historical interaction contexts, because the user's preferences may change over time. By considering similarity, time factor, and user satisfaction, etc. in multiple dimensions to select the best reference context. Optionally, the device implements a scenario clustering analysis mechanism to cluster and analyze the user's historical interaction records according to information type combinations, task goals, use scenarios, etc. features, identify the user's typical interaction patterns and preference types, then match the current context with each cluster center, select the most similar cluster as a reference, and extract the user's typical selection patterns and preference features from the cluster to generate personalized simplification decision suggestions. It can be understood that other ways of implementing scenario matching and analysis can also be used, which are not limited here. It needs to be supplemented that when the historical interaction records are insufficient or no enough similar historical contexts are found, the device will use a recommendation algorithm based on user group behavior patterns as a backup scheme.

[0089] S213、Determine the historical interaction context with the largest intersection and exceeding the preset scenario adaptation threshold as the target historical interaction context, and the historical simplified information template type corresponding to the target historical interaction context as the target simplified information template.

[0090] After the similarity calculation of the historical interaction context is completed, the device needs to select the most suitable reference context from them and determine the current simplification strategy based on the experience of the context. Specifically, the device sorts all the calculated intersection sizes, identifies the historical interaction context with the largest intersection, and verifies whether the intersection size exceeds the preset context adaptation threshold to ensure that the selected historical context has sufficient reference value. The preset context adaptation threshold is usually dynamically determined according to the size of the current information type tag set, for example, set to 50% of the current set size or a fixed value such as 3 tags. The device determines the historical interaction context that meets the threshold condition and has the largest intersection as the target historical interaction context, and if there are multiple contexts with the same maximum intersection, further considers factors such as time recency, user satisfaction score, etc. to select. The device extracts the simplification information template type used at the time from the record of the target historical interaction context, which may be "deletion type" (similar to the first simplification information template), "retention type" (similar to the second simplification information template), or other custom types. The device extracts the historical simplification information template type as the target simplification information template under the current circumstances, and adjusts the template adaptively according to the current specific information content and layout requirements to generate the final simplification scheme.

[0091] In some embodiments, the selection of the target historical interaction context and the determination of the template type can be implemented in various ways. Optionally, the device adopts a multi-factor comprehensive evaluation mechanism, in addition to the intersection size, also considers the time distance of the historical context, the user's satisfaction score at the time, the task completion efficiency, the subsequent use frequency, etc., calculates the comprehensive adaptation degree of each historical context through weighted scoring, selects the context with the highest comprehensive score as the target historical interaction context, and establishes a template adaptation verification mechanism to check whether the historical template type is suitable for the current information structure and user state, and adjusts and optimizes if necessary. Optionally, the device implements a dynamic threshold adjustment strategy, dynamically adjusts the preset context adaptation threshold according to the quality of the search results, appropriately reduces the threshold to expand the search range when there are fewer high-quality matching contexts, and increases the threshold to improve the selection accuracy when there are too many matching contexts, while adopting a template evolution mechanism to generate an improved version of the simplification template based on the characteristics of the historical template type and the current context, combining the advantages of historical experience and current needs. It can be understood that other ways can also be used to realize context selection and template determination, which are not limited here.

[0092] In the embodiments of the present application, since the technical scheme of binding specific information content units to negative emotion feature events, calculating the negative feedback rate of the information content units, and obtaining the content acceptance index by weighted summation based on the information weight, when the user has negative emotions, the specific information content that triggers the negative reaction can be located, the negative influence degree of different information units is quantitatively analyzed, the importance weight of the information units in the whole information is combined to calculate the comprehensive content acceptance evaluation, and then the targeted information simplification scheme is generated based on semantic understanding, and the suitable simplification strategy is intelligently selected through the natural gaze behavior or historical preference mode of the user, thereby effectively solving the technical problem of lack of personalized content optimization mechanism in the prior art, and achieving the technical effects of intelligent information presentation optimization and personalized cognitive burden adaptive adjustment, and improving the user experience quality and information acquisition efficiency of the AR glasses in the complex information display scene.

[0093] An example of an AR glass-based information adaptive presentation device 300 provided by the embodiments of the present application is introduced below. Figure 3 FIG. 1 is an example of a hardware structure schematic diagram of an AR glass-based information adaptive presentation device 300 provided by the embodiments of the present application.

[0094] In some embodiments, the AR glass-based information adaptive presentation device 300 is a computer device or includes a computer device in the AR glass-based information adaptive presentation device 300. The computer device includes a processor, a memory, and a network interface connected through a system bus. The processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for running the operating system and the computer program in the non-volatile storage medium. The database of the computer device is configured to store data. The network interface of the computer device is configured to communicate with other terminals or servers outside through a network connection. In some embodiments, the network interface can be a wired network interface, and in some embodiments, the network interface can also be a wireless network interface. The computer program is executed by the processor to implement the method in the embodiments of the present application.

[0095] Those skilled in the art can understand that Figure 3 The structure shown in FIG. 1 is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different component arrangement. The above-described embodiments merely serve to illustrate the technical solutions of the present application, rather than limit the present application; although the present application has been described in detail with reference to the foregoing embodiments, those ordinarily skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or equivalently replace some of the technical features thereof; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

[0096] In the above-described embodiments, the term "when" can be interpreted as meaning "if" or "after" or "in response to determining" or "in response to detecting" according to the context. Similarly, the phrase "upon determining" or "if detecting (the stated condition or event)" can be interpreted as meaning "if determining" or "in response to determining" or "upon detecting (the stated condition or event)" or "in response to detecting (the stated condition or event)" according to the context.

[0097] In the above-described embodiments, all or some of the flowcharts or functions can be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or some of the flowcharts or functions can be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When loaded and executed by a computer, all or some of the computer program instructions generate the flowcharts or functions described in the embodiments of the present application. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center through wired (such as coaxial cable, optical fiber, digital subscriber line) or wireless (such as infrared, wireless, microwave, etc.) mode. The computer-readable storage medium can be any available medium accessible by a computer or a data storage device such as a server, data center, etc. containing one or more available media. The available media can be magnetic media (such as floppy disks, hard disks, magnetic tapes), optical media (such as DVDs), or semiconductor media (such as solid state disks), etc.

[0098] Those of ordinary skill in the art can understand that all or part of the flowcharts in the above-described embodiments can be implemented by a computer program instructing relevant hardware, which can be stored in a computer-readable storage medium and executed to include the flowcharts of the above-described embodiments. The aforementioned storage medium includes ROM or random access memory (RAM), magnetic disks or optical disks, and various media that can store program codes.

Claims

1. An AR glasses-based information adaptive presentation method, characterized in that, Applied to AR glasses, the method comprises: At the beginning of the target user's current session, the ratio of the target user's total screen exposure time length to the average total screen exposure time length in the preset historical period is calculated based on the obtained cross-device usage data, and the ratio is taken as the daily accumulation index, the total screen exposure time length being the cumulative value of the target AR glasses usage time length and the total usage time length of the associated terminal device associated with the target user; In the case where the daily accumulation index is higher than the preset fatigue threshold, the facial micro-expression data of the target user in the preset time window are used to obtain a sequence of emotional feature events by combining an emotion recognition algorithm; The number of occurrences and the cumulative duration of negative emotional feature events in the emotional feature events are counted, and the content acceptance index is calculated by weighting based on the negative degree of the emotional feature events; When the content acceptance index is lower than the preset negative reaction threshold, the information simplification display switching operation is triggered.

2. The method of claim 1, wherein, After the step of calculating the ratio of the target user's total screen exposure time length to the average total screen exposure time length in the preset historical period based on the obtained cross-device usage data at the beginning of the target user's current session, the method further comprises: After the associated terminal device obtains the actual sleep time length, the average sleep demand time length, and the sleep quality score of the target user, a sleep recovery factor is obtained based on the ratio of the actual sleep time length to the average sleep demand time length multiplied by the sleep quality score; The ratio of the daily accumulation index to the sleep recovery factor is taken as the updated daily accumulation index.

3. The method of claim 1, wherein, The step of counting the number of occurrences and the cumulative duration of negative emotional feature events in the emotional feature events, and calculating the content acceptance index by weighting based on the negative degree of the emotional feature events, specifically comprises: The emotional feature event sequence is time-stamped aligned with the synchronous fixation point data obtained by eye tracking, and one or more information content units that the target user fixates on at the corresponding time are bound to each negative emotional feature event; The ratio of the cumulative duration of the negative emotional feature events associated with the information content unit to the total duration of the target user's fixation is calculated to obtain the negative feedback rate of the information content unit; The negative feedback rates are weighted and summed based on the information weight of the information content units in the current information panel to obtain the content acceptance index.

4. The method of claim 3, wherein, The step of triggering the information simplification display switching operation specifically comprises: Semantic analysis is performed on the currently displayed text content by calling a pre-trained natural language processing model to obtain a semantic analysis result, the semantic analysis including identifying information type labels and extracting key semantic entities, the information type labels including product description, technical specifications, and user feedback; Based on the semantic analysis result, at least two candidate simplified information templates are generated according to a preset information reorganization rule, the simplified information template being a simplified information formed by retaining the content corresponding to a specific information type label and compressing the content corresponding to a non-specific information type label; determine, based on the gaze duration of the target user on the at least two simplified information templates within a preset response time, a simplified information template with a gaze duration exceeding a preset selection threshold as a target simplified information template; perform the information simplification display switching operation based on the target simplified information template.

5. The method of claim 4, wherein, The step of generating at least two candidate simplified information templates according to a preset information reorganization rule based on the semantic analysis result specifically includes: After obtaining the negative feedback rate corresponding to all the information content units, determine the information type label with the highest negative feedback rate among all the information type labels as a high-negative-feedback information category and generate a first simplified information template, wherein the first simplified information template is formed by deleting the content belonging to the high-negative-feedback information category from the currently displayed page; determine, based on the negative feedback rate, an information category set with a negative feedback rate lower than a preset safety threshold as a low-negative-feedback information category set, and generate a second simplified information template, wherein the second simplified information template is formed by aggregating all the content belonging to the low-negative-feedback information category set; The first simplified information template and the second simplified information template are used as the candidate simplified information templates.

6. The method of claim 4, wherein, After the step of determining, based on the gaze duration of the target user on the at least two simplified information templates within a preset response time, a simplified information template with a gaze duration exceeding a preset selection threshold as a target simplified information template, the method further includes: In the case where it is determined that the target user does not respond within a preset response time, calculate, based on the information type label set contained in the currently displayed information, the intersection of the information type label in the historical interaction context within a preset time period and the current information type label set in the historical interaction record; determine a historical interaction context with the largest intersection and exceeding a preset scenario adaptation threshold as a target historical interaction context, and use the historical simplified information template type corresponding to the target historical interaction context as the target simplified information template.

7. The method of claim 1, wherein, Before the step of, in the case where the daily accumulation index is higher than a preset fatigue threshold, obtaining an emotion feature event sequence based on the facial micro-expression data of the target user within a preset time window and combining an emotion recognition algorithm, the method further includes: determine the difference between the average value and the standard deviation of the daily accumulation index of the target user within a preset statistical period in the past as the preset fatigue threshold.

8. An AR glasses based information adaptive presentation device, characterized by, The information adaptive presentation device based on AR glasses includes one or more processors and a memory; the memory is coupled with the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors invoke the computer instructions to make the information adaptive presentation device based on AR glasses execute the method in any one of claims 1-7.

9. A computer program product comprising instructions, characterized in that, When the computer program product runs on an information adaptive presentation device based on AR glasses, the information adaptive presentation device based on AR glasses executes the method in any one of claims 1-7. When the computer program product runs on an information adaptive presentation device based on AR glasses, the information adaptive presentation device based on AR glasses executes the method in any one of claims 1-7.

10. A computer-readable storage medium comprising instructions, characterized in that, When the instructions are run on an AR glasses based information adaptive presentation device, cause the AR glasses based information adaptive presentation device to perform the method of any of claims 1-7.

Images