Dynamic learning content generating and pushing method based on intelligent point reading system
User data is collected through multi-source sensors, the construction intensity and focus trajectory are constructed to detect the probability of distraction, and the learning content is dynamically adjusted, which solves the problem of disconnection between learning content and user needs in the intelligent point reading system, realizing personalized learning experience and interest stimulation.
Patent Information
- Application Number
- CN202510468776.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-15
- Publication Date
- 2025-07-18
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing intelligent point reading system cannot monitor the user's learning status in real time, resulting in the pushed learning content being disconnected from user needs, especially when distracted or inattentive, which cannot be adjusted in time.
Through multi-source sensors, users' usage intensity and focal focus data are collected in real time, the intensity change curve and focus movement trajectory are constructed, the probability of distraction is detected, and the appropriate learning content is selected from the structured knowledge base, and the teaching content is dynamically adjusted.
It realizes accurate monitoring of user attention, dynamically generates and pushes personalized learning content, enhances user interest and concentration, and provides personalized learning experience.
Smart Images
Figure CN120336635A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of computer technology, and particularly relates to a method for generating and pushing dynamic learning content based on an intelligent point-reading system. Background Art
[0002] Currently, the application of intelligent point-reading systems in the education field mainly focuses on the repetition of static content and simple interactions, lacking dynamic monitoring of users' learning states and adaptive adjustments. Traditional systems usually cannot capture the changes in users' attention in real time, resulting in the disconnection between the pushed learning content and users' actual needs. In addition, existing technologies mostly rely on a single sensor and are difficult to comprehensively evaluate users' learning states. Especially when users are distracted or inattentive, they cannot adjust the teaching content in time to re-attract users' attention. Therefore, there is an urgent need for a technical solution that can monitor users' learning states in real time, dynamically adjust teaching content, and improve learning effects. Summary of the Invention
[0003] In view of the above technical problems, the present invention provides a method for generating and pushing dynamic learning content based on an intelligent point-reading system. The intelligent point-reading system includes an intelligent point-reading machine, an intelligent point-reading pen, and a cloud server. The intelligent point-reading machine is used to connect to a point-reading textbook that allows users to learn. Users use the intelligent point-reading pen to click on the textbook content on the point-reading textbook connected to the intelligent point-reading machine to trigger the multimedia tool to repeat the clicked textbook content. The method is applied to the cloud server and includes: Real-time collecting user learning data of a target user through multi-source sensors. The user learning data includes the real-time usage force of the intelligent point-reading pen by the target user, the real-time focus coordinates of the target user's gaze focus, the target textbook content currently being learned by the target user, and the target point-reading textbook to which the target textbook content belongs; Constructing a force change curve of the target user according to the real-time usage force and the historical usage force of the target user, and constructing a focus movement trajectory of the target user according to the real-time focus coordinates and the historical focus coordinates of the target user. The historical usage force is the usage force of the intelligent point-reading pen by the target user within a historical time period, and the historical focus coordinates are the coordinates of the target user's gaze focus within a historical time period; Detecting the target distraction probability of the target user according to the force change curve and the gaze movement trajectory. The target distraction probability is used to indicate whether the attention of the target user has deviated from the target point-reading textbook; When the target distraction probability is greater than or equal to a distraction threshold, selecting target learning content from a structured knowledge base according to the target distraction probability and the user learning data. The structured knowledge base records the knowledge points corresponding to each textbook content in the target point-reading textbook, and records the association relationships between multiple knowledge points through a knowledge graph; Send the target learning content to the intelligent point-reading machine and the intelligent point-reading pen.
[0004] Further, a pressure sensor is deployed on the intelligent point-reading pen, and a camera is deployed on the intelligent point-reading machine. The user learning data of the target user is collected in real time through multi-source sensors, including: The holding pressure of the intelligent point-reading pen is collected in real time through the pressure sensor as the real-time usage force; The focus coordinates of the target user's gaze focus in the two-dimensional coordinate system are collected in real time through the camera as the real-time focus coordinates. The camera is used to construct a two-dimensional coordinate system with the intelligent point-reading machine as the plane and the camera as the coordinate origin; Obtain the target teaching material content and the target point-reading teaching material from the intelligent point-reading machine. The intelligent point-reading machine is used to identify the connected point-reading teaching material when the point-reading teaching material is connected, and to identify the teaching material content clicked by the user when it detects that the user uses the intelligent point-reading pen to click on the teaching material content.
[0005] Further, construct a force change curve of the target user according to the real-time usage force and the historical usage force of the target user, including: Obtain the usage force of the target user on the intelligent point-reading pen during the historical time period as the historical usage force; through Construct the force change curve F(t), , is the standard deviation of the historical usage force, T is the time length of the historical time period, F t is the real-time usage force, F history (t) is the historical usage force; Construct the focus movement trajectory of the target user according to the real-time focus coordinates and the historical focus coordinates of the target user, including: obtaining the coordinates of the gaze focus of the target user during the historical time period as the historical focus coordinates; arranging the historical focus coordinates and the real-time focus coordinates in the order from far to near in time; connecting the focus coordinates in the historical focus coordinates to the next focus coordinate in turn until the real-time focus coordinates are connected to obtain the focus movement trajectory.
[0006] Further, detect the target distraction probability of the target user according to the force change curve and the gaze movement trajectory, including: Detect the first proportion of the usage force deviating from the target force range in the force change curve in the force change curve. The target force range is 0.5 Newton to 3 Newtons; Detect the second proportion of the focus coordinates in the gaze movement trajectory whose distance from the previous focus coordinate is greater than or equal to the distance threshold in the gaze movement trajectory. The distance threshold is the length of the target point-reading teaching material; Calculate the weighted sum of the first proportion and the second proportion as the target distraction probability. The weight of the first proportion is 0.6, and the weight of the second proportion is 0.4.
[0007] Further, selecting target learning content from the structured knowledge base according to the target distraction probability and user learning data includes: Matching the target knowledge points corresponding to the target teaching materials content from the knowledge graph, and matching the knowledge point range corresponding to the target point-reading teaching materials from the knowledge graph; Detecting the content adaptation score between each knowledge point in the knowledge point range and the target knowledge point, and the content adaptation score is used to indicate the preference status of the target user for each knowledge point; Selecting the adaptation score interval falling into the target distraction probability from the corresponding relationship between the distraction probability and the adaptation score interval. The corresponding relationship between the distraction probability and the adaptation score interval includes that the adaptation score interval corresponding to the distraction probability greater than or equal to 0.85 is the first interval, the adaptation score interval corresponding to the distraction probability greater than or equal to 0.75 and less than 0.85 is the second interval, and the adaptation score interval corresponding to the distraction probability greater than or equal to 0.65 and less than 0.75 is the third interval, and the first interval is greater than the second interval and the second interval is greater than the third interval; Selecting the content adaptation scores falling into the target adaptation score interval, and extracting the knowledge points corresponding to each content adaptation score from the structured knowledge base as reference knowledge points; Selecting the intervention strategy corresponding to the target distraction probability from the corresponding relationship between the distraction probability and the intervention strategy as the target intervention strategy. The corresponding relationship between the distraction probability and the intervention strategy includes that the intervention strategy corresponding to the distraction probability greater than or equal to 0.85 is to display the knowledge points through animation, the intervention strategy corresponding to the distraction probability greater than or equal to 0.75 and less than 0.85 is to display the knowledge points through games, and the intervention strategy corresponding to the distraction probability greater than or equal to 0.65 and less than 0.75 is to display the knowledge points through exercises; Reconstructing the reference knowledge points using the target intervention strategy to obtain the target learning content.
[0008] Further, detecting the content adaptation score between each knowledge point in the knowledge point range and the target knowledge point includes: Traversing each knowledge point other than the target knowledge point in the knowledge point range in turn starting from the target knowledge point, taking the traversed knowledge point as a candidate knowledge point to execute the following steps until each knowledge point in the knowledge point range is traversed to obtain the corresponding relationship between the knowledge point and the content adaptation score: Calculating the semantic similarity between the candidate knowledge point and the target knowledge point, and the semantic similarity is positively correlated with the knowledge graph path distance; Obtaining the preference weight of the target user for the candidate knowledge point, and the preference weight is positively correlated with the number of times the target user learns the candidate knowledge point; Calculate the weighted sum of the computational semantic similarity and the preference weight as the content adaptation score of the candidate knowledge points.
[0009] The beneficial effects of the present invention compared with the prior art are as follows: (1) Accurately monitor the user's attention: By analyzing the user's usage force and the moving trajectory of the eye gaze focus, the distraction probability of the user can be accurately detected, providing a scientific basis for subsequent content push. (2) Dynamic content generation and push: According to the user's distraction probability and learning data, adaptively select appropriate learning content from the structured knowledge base, and reconstruct the content in the form of animations, games, or exercises, effectively improving the user's learning interest and concentration. (3) Personalized learning experience: Combine the knowledge graph and the user preference weight to generate knowledge points highly relevant to the user's current learning content, realizing true personalized learning. Brief Description of the Drawings
[0010] Figure 1 It is a flowchart of a method for generating and pushing dynamic learning content based on an intelligent point reading system of the present invention. Detailed Embodiments
[0011] Embodiment: As Figure 1 shown, a method for generating and pushing dynamic learning content based on an intelligent point reading system, the intelligent point reading system includes an intelligent point reading machine, an intelligent point reading pen, and a cloud server. The intelligent point reading machine is used to connect to the point reading textbooks that allow users to learn. The user uses the intelligent point reading pen to click on the textbook content on the point reading textbooks connected by the intelligent point reading machine to trigger the multimedia tool to repeat the clicked textbook content. The method is applied to the cloud server, and the method includes: Step S101: Real-time collect the user learning data of the target user through multi-source sensors. The user learning data includes the real-time usage force of the target user on the intelligent point reading pen, the real-time focus coordinates of the eye gaze focus of the target user, the target textbook content currently learned by the target user, and the target point reading textbook to which the target textbook content belongs; Step S101: Construct the force change curve of the target user according to the real-time usage force and the historical usage force of the target user, and construct the focus movement trajectory of the target user according to the real-time focus coordinates and the historical focus coordinates of the target user. The historical usage force is the usage force of the target user on the intelligent point reading pen within the historical time period, and the historical focus coordinates are the coordinates of the eye gaze focus of the target user within the historical time period; Step S102: Detect the target distraction probability of the target user according to the force change curve and the eye movement trajectory. The target distraction probability is used to indicate whether the attention of the target user has deviated from the target point reading textbook; Step S103: When the target distraction probability is greater than or equal to the distraction threshold, select target learning content from the structured knowledge base according to the target distraction probability and user learning data. The structured knowledge base records the knowledge points corresponding to each teaching material content in the target point-reading teaching material, and the association relationships between multiple knowledge points are recorded through a knowledge graph. Step S104: Send the target learning content to the intelligent point-reader and the intelligent point-reading pen.
[0012] In this embodiment, the point-reading teaching material refers to a physical or digital book that the user is learning (such as an English textbook, a math workbook, etc.), which is composed of specific teaching material content (such as words, formulas, paragraphs). The point-reading teaching material is used to provide structured learning materials (such as chapters, page numbers).
[0013] In this embodiment, when the user clicks on the teaching material content with the intelligent point-reading pen, it triggers a system feedback (such as voice reading). For example, a certain page in a primary school English textbook contains the word "apple", the sentence "This is an apple.", and related illustrations. When the user clicks on the word "apple" with the intelligent point-reading pen, it will wake up the voice device to read "apple".
[0014] In this embodiment, the knowledge point is the smallest knowledge unit corresponding to the teaching material content and is the basic unit for the system to make personalized recommendations.
[0015] In this embodiment, multiple knowledge points are associated through a knowledge graph. For example, "apple" is associated with knowledge points such as "fruit", "red", and "healthy". Knowledge points have semantic relationships (such as "addition" and "multiplication" belonging to the same branch of mathematics knowledge).
[0016] In this embodiment, when the user is distracted, the system matches content related to the current teaching material content from the knowledge point library (for example, when the user is distracted while reading "apple", it pushes knowledge points such as "banana" or "fruit classification"). For example, the knowledge point corresponding to the teaching material content "2 + 3 = 5" is "addition operation", and its associated knowledge points may include "number recognition", "carry addition", etc.
[0017] In this embodiment, the teaching material is the carrier of knowledge points, and the knowledge point is the abstract expression of the teaching material content. For example, a physics textbook contains multiple knowledge points such as "Newton's laws" and "gravity calculation". The system extracts associated knowledge points from the knowledge graph according to the teaching material content that the user is currently learning, and then dynamically pushes them in combination with the user's distraction level (such as animations, games, exercises).
[0018] In this embodiment, the same knowledge point can exist across different textbooks (e.g., "fraction operation" appears in both math textbooks and exercise books). The relevance of the knowledge graph allows the system to go beyond the current textbook and recommend interdisciplinary content (e.g., when learning the English word "apple", push the knowledge point of "plant fruits" in biology). For example: When a user uses a point-reading pen to learn the ancient poem "Thoughts in the Silent Night" in a primary school Chinese textbook, the textbook content is the line "Before my bed a pool of light". The direct knowledge points are the author of the ancient poem, "Li Bai", and the interpretation of the line. Extended knowledge points: Tang Dynasty literature and astronomical knowledge of the moon (associated through the knowledge graph). Distraction intervention: If the probability of the user being distracted is high, the system may push an animation about "Li Bai's life" or an interactive game about "moon phases". Through this design, the system realizes the leap from fixed textbook content to flexible knowledge point recommendation, enhancing the personalization and interest of learning.
[0019] Optionally, a pressure sensor is deployed on the intelligent point-reading pen, and a camera is deployed on the intelligent point-reading machine. User learning data of the target user is collected in real time through multi-source sensors, including: the holding pressure of the intelligent point-reading pen is collected in real time through the pressure sensor as the real-time usage force; the focus coordinates of the target user's eye focus in the two-dimensional coordinate system are collected in real time through the camera as the real-time focus coordinates. The camera is used to construct a two-dimensional coordinate system with the intelligent point-reading machine as the plane and the camera as the coordinate origin; the target textbook content and the target point-reading textbook are obtained from the intelligent point-reading machine. The intelligent point-reading machine is used to identify the connected point-reading textbook when the point-reading textbook is connected, and to identify the textbook content clicked by the user when it detects that the user uses the intelligent point-reading pen to click on the textbook content.
[0020] In this embodiment, the intelligent point-reading machine can, but is not limited to, identify the connected point-reading textbook in multiple ways, such as: (1) Physical identification detection: QR code / RFID tag: A unique identification code is embedded on the cover of each textbook, and the point-reading machine automatically identifies information such as the textbook name, version, and subject by scanning. ISBN code recognition: The ISBN number of the textbook is read through OCR technology and matched with the cloud database to obtain textbook metadata.
[0021] (2) Image feature matching: Cover / binding recognition: The textbook cover is photographed using the camera, and features are extracted through a pre-trained CNN model (such as ResNet) and compared with the textbook cover features in the database. Page layout analysis: Detect the fixed layout of the textbook inner pages (such as the font of the header and chapter titles) to assist in confirming the textbook type.
[0022] (3) If the identification is damaged, the user can manually select from the textbooks cached in the system.
[0023] In this embodiment, the intelligent point reader can, but is not limited to, identify the teaching material content clicked by the user in multiple ways when detecting that the user uses the intelligent point pen to click on the teaching material content. For example: (1) Coordinate mapping method: The electronic version of each teaching material pre-stores a coordinate-content mapping table, and the click position (x, y) of the point pen triggers a query: (2) OCR real-time recognition: The camera of the point reader captures the area clicked by the user, and identifies text or formulas through OCR (such as Tesseract), and performs full-text retrieval and matching with the electronic version of the teaching material.
[0024] Multimodal fusion: Combine text recognition + image features (such as formula symbols, illustrations) to improve the accuracy.
[0025] Optionally, construct a force change curve of the target user according to the real-time usage force and the historical usage force of the target user, including: obtaining the usage force of the intelligent point pen by the target user within the historical time period as the historical usage force; by Construct a force change curve F(t), , is the standard deviation of the historical usage force, T is the time length of the historical time period, F t is the real-time usage force, F history (t) is the historical usage force; construct a focus movement trajectory of the target user according to the real-time focus coordinates and the historical focus coordinates of the target user, including: obtaining the coordinates of the gaze focus of the target user within the historical time period as the historical focus coordinates; arranging the historical focus coordinates and the real-time focus coordinates in the order from far to near in time; connecting the focus coordinates in the historical focus coordinates to the next focus coordinate in turn until the real-time focus coordinates are connected to obtain the focus movement trajectory.
[0026] Optionally, detect the target distraction probability of the target user according to the force change curve and the gaze movement trajectory, including: detecting the first proportion of the usage force deviating from the target force range in the force change curve in the force change curve, and the target force range is 0.5 Newton to 3 Newtons; detecting the second proportion of the focus coordinates with a distance greater than or equal to the distance threshold from the previous focus coordinate in the gaze movement trajectory in the gaze movement trajectory, and the distance threshold is the length of the target point-reading teaching material; calculating the weighted sum of the first proportion and the second proportion as the target distraction probability, and the weight of the first proportion is 0.6 and the weight of the second proportion is 0.4.
[0027] In this embodiment, the target force range can be updated in the following ways, but not limited to: (1) Dynamic calibration based on user historical data: Initial default range: The system preset basic force range (such as 0.5N - 3N), which is applicable to new users or when there is no historical data. (2) Personalized adjustment: Mean ± standard deviation method: Record the pen - holding force data of the user in the past 7 days, calculate the mean (μ) and standard deviation (σ), and set the target range as μ ± 0.5σ. Example: If μ = 2N and σ = 0.8N, then the range is 1.6N - 2.4N. (3) Percentile method: Take the 25% - 75% quantiles of the user's historical force data as the range (excluding extreme values).
[0028] Optionally, select the target learning content from the structured knowledge base according to the target distraction probability and user learning data, including: matching the target knowledge points corresponding to the target teaching materials content from the knowledge graph, and matching the knowledge point range corresponding to the target point - reading teaching materials from the knowledge graph; detecting the content adaptation score between each knowledge point in the knowledge point range and the target knowledge point, where the content adaptation score is used to indicate the preference status of the target user for each knowledge point; selecting the adaptation score interval that falls into the adaptation score interval corresponding to the target distraction probability from the corresponding relationship between the distraction probability and the adaptation score interval. The corresponding relationship between the distraction probability and the adaptation score interval includes that the adaptation score interval corresponding to the distraction probability greater than or equal to 0.85 is the first interval, the adaptation score interval corresponding to the distraction probability greater than or equal to 0.75 and less than 0.85 is the second interval, and the adaptation score interval corresponding to the distraction probability greater than or equal to 0.65 and less than 0.75 is the third interval, and the first interval is greater than the second interval and the second interval is greater than the third interval; select the content adaptation scores that fall into the target adaptation score interval, and extract the knowledge points corresponding to each content adaptation score from the structured knowledge base as reference knowledge points; select the intervention strategy corresponding to the target distraction probability from the corresponding relationship between the distraction probability and the intervention strategy as the target intervention strategy. The corresponding relationship between the distraction probability and the intervention strategy includes that the intervention strategy corresponding to the distraction probability greater than or equal to 0.85 is to display the knowledge points through animation, the intervention strategy corresponding to the distraction probability greater than or equal to 0.75 and less than 0.85 is to display the knowledge points through games, and the intervention strategy corresponding to the distraction probability greater than or equal to 0.65 and less than 0.75 is to display the knowledge points through exercises; reconstruct the reference knowledge points using the target intervention strategy to obtain the target learning content.
[0029] In this embodiment, when the distraction probability is greater than or equal to 0.85, it is considered that the user's state is significantly distracted (such as frequently lowering the head, the pen - holding force continuously being lower than 0.3N). Therefore, the first interval is set as the high - intervention area, and it is necessary to select knowledge points with a high overlap (high semantic similarity) with the knowledge points that the user is currently learning and that the user is interested in (high preference weight), and the intervention strategy is to attract attention with high - intensity dynamic content.
[0030] When the distraction probability is greater than or equal to 0.75 and less than 0.85, the user's state is considered to be mildly distracted (such as the gaze shifting but the force being normal). Therefore, the second interval is the medium intervention area, and it is necessary to balance the selection of knowledge points that are similar to the knowledge points the user is currently learning (average semantic similarity) and that the user is more interested in (average preference weight). The intervention strategy is to guide the regression with medium-interactivity content.
[0031] When the distraction probability is greater than or equal to 0.65 and less than 0.75, the user's state is considered to be potentially distracted (such as occasional eye wandering). Therefore, the third interval is the low intervention area, and knowledge points that are relatively different from the knowledge points the user is currently learning (low semantic similarity) or that the user has not paid attention to (low preference weight) can be appropriately selected. The intervention strategy is to consolidate concentration with lightweight prompts.
[0032] In this embodiment, the range obtained by detecting the content adaptation scores between each knowledge point in the knowledge point range and the target knowledge point can be evenly divided into three parts as the first interval, the second interval, and the third interval of the above adaptation score interval.
[0033] It should be noted that the corresponding relationship between the magnitude of the distraction probability and the intervention strategy is determined according to the distraction threshold, and the distraction threshold in the above embodiment is 0.65.
[0034] Optionally, detecting the content adaptation scores between each knowledge point in the knowledge point range and the target knowledge point includes: starting from the target knowledge point, traversing each knowledge point in the knowledge point range except the target knowledge point in turn, taking the traversed knowledge point as a candidate knowledge point to perform the following steps until each knowledge point in the knowledge point range is traversed to obtain the corresponding knowledge point and content adaptation score: calculating the semantic similarity between the candidate knowledge point and the target knowledge point, where the semantic similarity is positively correlated with the knowledge graph path distance; obtaining the preference weight of the target user for the candidate knowledge point, where the preference weight is positively correlated with the number of times the target user learns the candidate knowledge point; calculating the weighted sum of the semantic similarity and the preference weight as the content adaptation score of the candidate knowledge point.
[0035] In the technical solution proposed by the present invention, the present invention provides a method for generating and pushing dynamic learning content based on an intelligent point-reading system. Through multi-dimensional data collection, intelligent analysis, and personalized content pushing, real-time monitoring and dynamic optimization of the learning process are achieved, which has significant technological innovation and practical value.
[0036] In terms of technical effects, the present invention first collects the user's grip strength and gaze focus data in real time through pressure sensors and cameras, and constructs a strength change curve and a focus movement trajectory in combination with historical learning behaviors, which can accurately identify the user's distracted state. Compared with traditional point-reading systems that only rely on simple interaction records, the present invention uses comprehensive analysis of multi-source sensor data, significantly improving the accuracy and real-time performance of attention detection.
[0037] Secondly, the system associates knowledge points through a knowledge graph and dynamically adjusts the pushed content in combination with the user's distraction probability. When the user's attention drops, the system can intelligently select suitable learning materials according to the degree of distraction and reconstruct the content using differential strategies such as animations, games, or exercises, effectively stimulating the learning interest. This mechanism not only solves the problems of single and lack of personalization in the pushed content of traditional point-reading systems, but also realizes the dynamic matching of learning content and user status, thus greatly improving the learning efficiency and knowledge retention rate.
[0038] In addition, the present invention realizes real-time data processing and feedback through the collaborative work of the cloud server and the intelligent point-reading device, ensuring the response speed and stability of the system. Users can obtain personalized learning support without awareness, and educators can also optimize teaching strategies through system feedback.
[0039] In summary, the present invention significantly improves the interactivity and teaching effect of the point-reading system through an intelligent and dynamic learning content generation and pushing mechanism, provides an efficient and accurate solution for the field of intelligent education, and has broad application prospects.
[0040] In the description of this specification, the descriptions referring to terms such as "one embodiment", "example", "specific example", etc. mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example.
Claims
1. A method for generating and pushing dynamic learning content based on an intelligent point-reading system, characterized in that, The intelligent point-reading system includes an intelligent point-reading machine, an intelligent point-reading pen, and a cloud server. The intelligent point-reading machine is used to connect to the point-reading textbooks that allow users to learn. The user uses the intelligent point-reading pen to click on the textbook content on the point-reading textbook connected to the intelligent point-reading machine to trigger the multimedia tool to repeat the clicked textbook content. The method is applied to the cloud server, and the method includes: Real-time collecting user learning data of the target user through multi-source sensors. The user learning data includes the real-time usage force of the intelligent point-reading pen by the target user, the real-time focus coordinates of the target user's eye gaze focus, the target textbook content currently being learned by the target user, and the target point-reading textbook to which the target textbook content belongs; Constructing a force change curve of the target user according to the real-time usage force and the historical usage force of the target user, and constructing a focus movement trajectory of the target user according to the real-time focus coordinates and the historical focus coordinates of the target user. The historical usage force is the usage force of the intelligent point-reading pen by the target user within a historical time period, and the historical focus coordinates are the coordinates of the target user's eye gaze focus within the historical time period; Detecting the target distraction probability of the target user according to the force change curve and the eye movement trajectory. The target distraction probability is used to indicate whether the attention of the target user has deviated from the target point-reading textbook; When the target distraction probability is greater than or equal to the distraction threshold, selecting target learning content from the structured knowledge base according to the target distraction probability and the user learning data. The structured knowledge base records the knowledge points corresponding to each textbook content in the target point-reading textbook, and records the association relationships between multiple knowledge points through a knowledge graph; Sending the target learning content to the intelligent point-reading machine and the intelligent point-reading pen.
2. The dynamic learning content generation and pushing method based on an intelligent point-reading system according to claim 1, wherein A pressure sensor is deployed on the intelligent point-reading pen, and a camera is deployed on the intelligent point-reading machine. The real-time collecting user learning data of the target user through multi-source sensors includes: Real-time collecting the holding pressure of the intelligent point-reading pen by the pressure sensor as the real-time usage force; Real-time collecting the focus coordinates of the target user's eye gaze focus in the two-dimensional coordinate system by the camera as the real-time focus coordinates. The camera is used to construct the two-dimensional coordinate system with the intelligent point-reading machine as the plane and the camera as the coordinate origin; Obtaining the target textbook content and the target point-reading textbook from the intelligent point-reading machine. The intelligent point-reading machine is used to identify the connected point-reading textbook when it is connected to the point-reading textbook, and to identify the textbook content clicked by the user when it detects that the user uses the intelligent point-reading pen to click on the textbook content.
3. The dynamic learning content generation and pushing method based on an intelligent point-reading system according to claim 1, characterized in that, The constructing a force change curve of the target user according to the real-time usage force and the historical usage force of the target user includes: Obtain the usage intensity of the intelligent point-reading pen by the target user during the historical time period as the historical usage intensity; By Construct the intensity change curve F(t), , is the standard deviation of the historical usage intensity, T is the time length of the historical time period, F t is the real-time usage intensity, F history (t) is the historical usage intensity; Constructing the focus movement trajectory of the target user according to the real-time focus coordinates and the historical focus coordinates of the target user includes: obtaining the coordinates of the gaze focus of the target user within the historical time period as the historical focus coordinates; arranging the historical focus coordinates and the real-time focus coordinates in the order from far to near in time; and connecting the focus coordinates in the historical focus coordinates to the next focus coordinate in sequence until the real-time focus coordinates are connected to obtain the focus movement trajectory.
4. A method for generating and pushing dynamic learning content based on an intelligent point-reading system according to claim 1, characterized in that Detecting the target distraction probability of the target user according to the force change curve and the gaze movement trajectory includes: Detecting a first proportion of the usage force deviating from the target force range in the force change curve, where the target force range is from 0.5 Newton to 3 Newtons; Detecting a second proportion of the focus coordinates in the gaze movement trajectory whose distance from the previous focus coordinate is greater than or equal to the distance threshold, where the distance threshold is the length of the target point-reading textbook; Calculating the weighted sum of the first proportion and the second proportion as the target distraction probability, where the weight of the first proportion is 0.6 and the weight of the second proportion is 0.
4.
5. The dynamic learning content generation and pushing method based on an intelligent point-reading system according to claim 1, characterized in that, Selecting the target learning content from the structured knowledge base according to the target distraction probability and the user learning data includes: Matching the target knowledge points corresponding to the target textbook content from the knowledge graph, and matching the knowledge point range corresponding to the target point-reading textbook from the knowledge graph; Detecting the content adaptation score between each knowledge point in the knowledge point range and the target knowledge point, where the content adaptation score is used to indicate the preference state of the target user for each knowledge point; Selecting the adaptation score interval that falls into the adaptation score interval corresponding to the target distraction probability from the corresponding distraction probability and adaptation score interval, where the corresponding distraction probability and adaptation score interval include that the adaptation score interval corresponding to the distraction probability greater than or equal to 0.85 is the first interval, the adaptation score interval corresponding to the distraction probability greater than or equal to 0.75 and less than 0.85 is the second interval, and the adaptation score interval corresponding to the distraction probability greater than or equal to 0.65 and less than 0.75 is the third interval, and the first interval is greater than the second interval and the second interval is greater than the third interval; Selecting the content adaptation scores that fall into the target adaptation score interval, and extracting the knowledge points corresponding to each content adaptation score from the structured knowledge base as reference knowledge points; Selecting the intervention strategy corresponding to the target distraction probability from the corresponding distraction probability and intervention strategy as the target intervention strategy, where the corresponding relationship between the distraction probability and the intervention strategy includes that the intervention strategy corresponding to the distraction probability greater than or equal to 0.85 is to display the knowledge points through animation, the intervention strategy corresponding to the distraction probability greater than or equal to 0.75 and less than 0.85 is to display the knowledge points through games, and the intervention strategy corresponding to the distraction probability greater than or equal to 0.65 and less than 0.75 is to display the knowledge points through exercises; Reconstruct the reference knowledge points using the target intervention strategy to obtain the target learning content.
6. The method for generating and pushing dynamic learning content based on an intelligent point-reading system according to claim 5, wherein The detection of the content adaptation scores between each knowledge point in the knowledge point range and the target knowledge point includes: Starting from the target knowledge point, traverse each knowledge point in the knowledge point range except the target knowledge point in turn. Take the traversed knowledge point as a candidate knowledge point and execute the following steps until each knowledge point in the knowledge point range is traversed to obtain the corresponding knowledge point and content adaptation score: Calculate the semantic similarity between the candidate knowledge point and the target knowledge point, and the semantic similarity is positively correlated with the knowledge graph path distance; Obtain the preference weight of the target user for the candidate knowledge point, and the preference weight is positively correlated with the number of times the target user learns the candidate knowledge point; Calculate the weighted sum of the semantic similarity and the preference weight as the content adaptation score of the candidate knowledge point.