Method, device and equipment for selecting point reading conflict book and storage medium

Abstract

The invention relates to the technical field of deep learning, and discloses a method, a device and equipment for selecting a touch-and-read conflict book and a storage medium, and the method comprises the steps that the touch-and-read equipment collects time and position data of touch-and-read operation of a user, transmits the time and position data to a server, and generates a touch-and-read behavior feature record of the user; judging whether a conflict condition that multiple books correspond to the same point reading code exists or not, and sending conflict prompt information to the mobile terminal when the conflict exists; generating a conflict book sorting list, and sending voice prompt information to a point reading device to guide a user to select books on a mobile terminal; a user selects a target book on the mobile terminal, establishes a binding relationship and stores the binding relationship in a user configuration database; when the user subsequently uses the same click-reading code to perform click-reading, the audio resource file of the target book is automatically returned, the user does not need to select again, a cooperative working mechanism of the click-reading device and the mobile terminal is realized, and the accuracy and convenience of conflict book selection are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of deep learning, and in particular, to a method, device, equipment and storage medium for selecting conflict books for point reading. Background Art

[0002] With the in-depth development of educational informatization, the point reading pen has become an important tool for early childhood education. Especially in the stage when children cannot fully recognize words, the point reading pen is widely used to assist knowledge acquisition due to its intuitive and convenient features. With the progress of technology, intelligent WIFI point reading pens have emerged. By downloading book audio resources to the pen through the network, the point reading experience becomes more fluent and resource acquisition becomes more convenient, so they have been widely welcomed in the market. However, with the popularization of point reading pens and the expansion of their application scope, a technical problem has become increasingly prominent.

[0003] Due to the fixed limit of the number of point reading code values, and the large number of publishing houses, the number of book resources has grown explosively. Coupled with the fact that audio resources in modern point reading systems are mainly stored on the server side, there will inevitably be a conflict situation where multiple different books share the same point reading code. This code value conflict directly causes the point reading pen to be unable to determine which specific book the user is using during the recognition process, and thus unable to provide the correct audio feedback, seriously affecting the normal use experience of users. Existing solutions usually require users to contact customer service. After the customer service confirms the book used by the user, the audio resources are sent to the user by the offline file transfer method, and then the user downloads them to the point reading pen through the computer USB interface. Summary of the Invention

[0004] The present invention provides a method, device, equipment and storage medium for selecting conflict books for point reading. The present invention realizes the collaborative working mechanism between the point reading device and the mobile terminal, and improves the accuracy and convenience of selecting conflict books.

[0005] In the first aspect, the present invention provides a method for selecting conflict books for point reading. The method for selecting conflict books for point reading includes: The point reading device collects the time and position data of the user's point reading operation and transmits them to the server to generate a record of the user's point reading behavior characteristics; The server receives the point reading code sent by the point reading device and determines whether there is a conflict situation where multiple books correspond to the same point reading code. When there is a conflict, it sends a conflict prompt message to the mobile terminal; The server calculates the matching degree between each conflict book and the current point reading situation according to the record of the user's point reading behavior characteristics, and generates a sorted list of conflict books; The server sends the sorted list of conflict books to the mobile terminal, and at the same time sends a voice prompt message to the point reading device to guide the user to select a book on the mobile terminal; After the user selects a target book on the mobile terminal, the server receives the selection result, establishes a binding relationship between the target book selected by the user and the point-reading code, and stores it in the user configuration database; According to the binding relationship, when the user subsequently uses the same point-reading code for point-reading, the server automatically returns the audio resource file of the target book without the user having to select again.

[0006] In a second aspect, the present invention provides a device for selecting books with point-reading conflicts, and the device for selecting books with point-reading conflicts includes: An acquisition module, configured to collect the time and position data of the user's point-reading operation by the point-reading device and transmit it to the server, and generate a record of the user's point-reading behavior characteristics; A judgment module, configured to receive the point-reading code sent by the point-reading device by the server and judge whether there is a conflict situation where multiple books correspond to the same point-reading code. When there is a conflict, a conflict prompt message is sent to the mobile terminal; A calculation module, configured to calculate the matching degree between each conflicting book and the current point-reading situation according to the record of the user's point-reading behavior characteristics by the server, and generate a sorted list of conflicting books; A guidance module, configured to send the sorted list of conflicting books to the mobile terminal by the server, and at the same time send a voice prompt message to the point-reading device to guide the user to select a book on the mobile terminal; An establishment module, configured to after the user selects a target book on the mobile terminal, the server receives the selection result, establishes a binding relationship between the target book selected by the user and the point-reading code, and stores it in the user configuration database; An automatic return module, configured to according to the binding relationship, when the user subsequently uses the same point-reading code for point-reading, the server automatically returns the audio resource file of the target book without the user having to select again.

[0007] In a third aspect of the present invention, there is provided a computer device, including: a memory and at least one processor, wherein instructions are stored in the memory; the at least one processor calls the instructions in the memory to enable the computer device to execute the method for selecting books with point-reading conflicts as described above.

[0008] In a fourth aspect of the present invention, there is provided a computer-readable storage medium, wherein instructions are stored in the computer-readable storage medium, and when it runs on a computer, it enables the computer to execute the method for selecting books with point-reading conflicts as described above.

[0009] In the technical solution provided by the present invention, the server automatically detects the point-reading code conflict situation and sends prompt messages to the point-reading device and the mobile phone terminal simultaneously, realizing the timely discovery and active notification of conflicts, avoiding user confusion and annoyance, and clearly guiding users to take subsequent operations. By using the user point-reading behavior data collected by the point-reading device, a user point-reading context relationship graph is constructed. By analyzing the user's point-reading habits and content preferences, the matching degree between the conflicting book and the current point-reading situation is intelligently calculated, providing personalized recommendation sorting for users, and greatly improving the accuracy and convenience of selecting conflicting books. A collaborative working mechanism between the point-reading device and the mobile terminal is realized. By combining the voice prompt of the point-reading device and the visual interface of the mobile terminal, an intuitive and clear book selection process is created, reducing the user operation threshold and enhancing the interaction experience. The corresponding relationship of the user's book selection result is established and stored in the user configuration database, realizing the function of "select once, remember permanently". When the user uses the same point-reading code later, the system automatically returns the audio resources of the selected book without repeating the selection process, greatly simplifying the user operation steps. By analyzing the user's historical selection behavior and point-reading habits, a dynamic weight mechanism is established for conflicting books, and a higher priority is given to the content types preferred by users, continuously optimizing the user point-reading experience and making the system more in line with the personalized needs of users. The present invention completely realizes the solution of point-reading conflicts and the acquisition of audio resources through the network, avoiding the cumbersome operation of downloading audio resources through USB in the traditional solution, eliminating the dependence on personal computers, enabling all users to conveniently solve the point-reading conflict problem, and improving the universality and user-friendliness of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0011] Figure 1 It is a schematic step diagram of the method for selecting a point-reading conflict book in an embodiment of the present invention; Figure 2 It is a schematic structural diagram of the device for selecting a point-reading conflict book in an embodiment of the present invention; Figure 3 It is a schematic block diagram of the structure of a computer device in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0012] An embodiment of the present invention provides a method, apparatus, device, and storage medium for selecting point-reading conflicting books. The terms "first", "second", "third", "fourth", etc. (if any) in the specification, claims, and the above-mentioned drawings of the present invention are used to distinguish similar objects and do not necessarily describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments described herein can be implemented in an order other than those illustrated or described herein. In addition, the term "comprising" or "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0013] For ease of understanding, the specific process of the embodiment of the present invention is described below. Please refer to Figure 1 In an embodiment of the method for selecting point-reading conflicting books in the embodiment of the present invention, it includes: Step S1: The point-reading device collects the time and position data of the user's point-reading operation and transmits it to the server to generate a record of the user's point-reading behavior characteristics; It can be understood that the execution subject of the present invention can be a device for selecting point-reading conflicting books, or a terminal or a server. Specifically, it is not limited here. The embodiment of the present invention is described by taking the server as the execution subject as an example.

[0014] Specifically, the time and position data of the user's dot-reading operation are collected by the dot-reading device and transmitted to the server to generate a record of the user's dot-reading behavior characteristics. A high-precision position perception and time synchronization module is integrated at the dot-reading device end, and a series of behavior information generated by the user during dot-reading is continuously monitored and captured relying on the built-in sensors. At the moment when the user performs a dot-reading operation, the dot-reading device captures the two-dimensional coordinate position of the contact point between the dot-reading pen tip and the paper surface through infrared scanning, capacitive array or image recognition technology, supplements the pressure sensor to detect the strength level of the dot-reading pressure, and at the same time combines the internal real-time clock chip to record the specific time point when the dot-reading occurs and the duration of the dot-reading. The above multi-dimensional information constitutes the original dot-reading data set, and each record in this data set contains the time stamp when the dot-reading occurs, the corresponding coordinate value, the applied pressure intensity, and the time interval during which the operation lasts. The original dot-reading data is transmitted to the server, and format standardization and integrity verification are performed on the original dot-reading data. After passing the verification, a time series sorting operation is performed on all dot-reading records based on the time stamp to generate structured dot-reading behavior trajectory data with time series characteristics. According to the structured dot-reading time series data, pattern recognition processing based on the change of the spatial trajectory is performed. The change vector of the coordinates between two adjacent dot-reading operations is identified, and its direction angle and spatial distance in the two-dimensional plane are calculated. Then, the change rules of multiple consecutive point pairs are comprehensively evaluated to extract the main trend of the dot-reading direction of the user during the current period (such as from left to right, from top to bottom, etc.). At the same time, by comparing the time interval and stay duration between each dot-reading, the rhythm characteristics in the user's dot-reading process are summarized, such as rapid browsing, equal-interval reading or discontinuous jump scanning, etc. The system establishes a behavior feature vector on this basis, in which the directionality and rhythm are quantitatively expressed as two core parameters, and are fused with the user's historical behavior model to generate a record of the user's dot-reading behavior characteristics in the current context.

[0015] Step S2: The server receives the dot-reading code sent by the dot-reading device and determines whether there is a conflict situation where multiple books correspond to the same dot-reading code. When there is a conflict, a conflict prompt message is sent to the mobile terminal. Specifically, the device packs the currently collected point reading code value together with additional information such as the device identifier and the point reading timestamp to form a structured request data packet, and sends it to the point reading code processing module on the server side through the network. After receiving the request, the server extracts the point reading code from the data packet and uses it as the primary key to perform a query operation in the point reading code mapping database. All the associated mapping relationships between all standard point reading codes and their corresponding book resources are pre-stored in this database, covering metadata fields such as the unique ID of the book, the publisher, the book title, the page number structure, and the suitable age range for reading. Through matching retrieval, a set of matching results is obtained, and all candidate book entries that are the same as the current point reading code are listed in this set. The server then counts the total number of candidate books in the matching results and makes a judgment based on the set single correspondence threshold (usually 1). If the number of candidate books exceeds this threshold, it is preliminarily determined that there is a point reading code conflict, and the system generates a conflict detection result marked as "conflict exists". Thereafter, the conflict analysis module is called to compare the metadata information of all conflicting books in this result set, analyze the degree of difference between them, especially conduct a quantitative comparison from multiple dimensions such as book category (such as picture books, story collections, encyclopedias), content structure (such as picture-text ratio, page layout), publishing unit, and suitable age grading, and calculate the similarity score between each pair of books. According to the preset similarity threshold standard, the conflict situation is divided into three categories: low-level conflict, medium-level conflict, and high-level conflict. Among them, a low-level conflict indicates that the uses or target audiences of the books are significantly different, a medium-level conflict indicates that the appearances or structures of the books are different but the content themes are similar, and a high-level conflict indicates that the content and form are highly consistent and easy to be confused. Finally, a conflict type determination result is output. Based on this, the server generates a structured formatted conflict message, which is encapsulated in the standard JSON or XML format. The fields included are the number of conflicting books, the conflict level, the cover thumbnail link of each book, the book title, the publisher, the point reading code value itself, the user's current request context information, etc. At the same time, a unique session identifier assigned by the server is embedded, and this identifier is used to subsequently bind the user's selection operation to the current conflict handling session. After the formatted conflict message is generated, the server calls the message push service interface and immediately pushes this message to the bound user mobile terminal through the WebSocket long connection or the mobile push platform (such as FCM or APNs), triggering a pop-up reminder or a notification bar reminder on the mobile phone side, guiding the user to enter the conflict resolution interface, and the user performs a selection operation.

[0016] Step S3: The server calculates the matching degree of each conflicting book with the current point reading situation according to the user's point reading behavior characteristics record, and generates a sorted list of conflicting books; Specifically, through the behavior feature analysis module on the server side, key feature items representing the user's behavior pattern are extracted from the current user's point-reading behavior feature record. This process is based on the structured data of the behavior record, and extracts the book type tags of the user's recent consecutive point-reads (such as fairy tales, knowledge, language learning, etc.), typical point-reading time patterns (including the time period and frequency of point-reading), the distribution density of the point-reading position on the page (such as the habit of swiping from the upper left to the lower right of the page), and the book tags that the user tends to select in conflict situations in historical selection behavior. These feature items are encoded into numerical forms to construct the feature vector of the current point-reading situation. At the same time, the server performs feature extraction operations on all conflicting books corresponding to the current point-reading code. The extracted fields include the publication information of the book, content structure tags, suitable reading age range, graphic and text layout ratio, distribution of typical pattern positions, etc., and performs feature vectorization processing on each book to generate a set of book feature vectors with consistent structures. The server enters the similarity calculation stage, and evaluates the matching degree between the current point-reading situation feature vector and each book's feature vector respectively. The algorithm used is a multi-factor weighted Bayesian calculation model. In the implementation process of this model, weight coefficients are set for each behavior and book feature factor respectively, and the posterior probability of each conflicting book appearing in the current point-reading context is calculated according to Bayes' theorem, and this is used as the initial matching score. After the preliminary matching is completed, the user's historical selection preferences are introduced to dynamically adjust the matching results. If the system recognizes that the user has selected a specific publisher or book series multiple times under the same code value conflict, the target matching degree score of this book is increased; if the current matching degree is low but the user has a similar behavior trajectory in the past similar time period, its score is also appropriately increased. All conflicting books are sorted in descending order according to the target matching degree score to generate a structured list of sorted conflicting books, and this list is stored in a cache queue with time identification and session identification for subsequent pushing to the mobile terminal interface for display.

[0017] Step S4: The server sends the sorted list of conflicting books to the mobile terminal, and at the same time sends a voice prompt message to the point-reading device to guide the user to select a book on the mobile terminal; Specifically, after the server completes the sorting of the conflict book matching degrees, it calls the front-end interface generation module to structurally layout the obtained conflict book sorted list, perform position optimization processing on the sorting results according to the matching degree scores, set the recommended book with the highest score as the preferred item and place it in the top area of the user interface. In this area, elements such as the cover image of the recommended book, the book title, the publishing house information, and the age suitability prompt are prominently displayed to attract the user's attention and guide them to refer to it first; the remaining conflict books with lower matching degrees are arranged in the lower area of the interface in descending order of matching degree, forming a clear visual layering logic to enhance the user's judgment efficiency when facing multiple options. The server constructs a user interface display template for mobile terminal display based on this layout structure. This template adopts a responsive layout to adapt to different screen sizes and presets interactive response hooks for subsequent binding of user selection operations. After the interface display template is constructed, the server compresses all the book cover images involved in the template, using the JPEG compression algorithm or the WebP format to reduce the transmitted data volume while ensuring clarity. At the same time, a unique selection button, confirmation button, and "set as default" operation entry are attached to each book image to complete the binding of the interactive elements of the interface template. The generated interactive selection interface includes a layout combining text and graphics, a clearly divided operation area, and a complete event response mechanism. The server encapsulates this interface in structured data (such as JSON) and pushes it to the point reading pen supporting application on the user's mobile phone in real time through a WebSocket or HTTP long connection channel to achieve instant loading and display of the interface. At the same time, the server synchronously starts the voice prompt instruction generation module of the point reading device and automatically constructs standard voice content according to the currently detected conflict type and sorting results. The system calls the text-to-speech engine to convert the prompt into voice waveform data and encodes it into a control instruction structure that conforms to the recognition standard of the point reading device to form a voice playback control data packet. This data packet includes voice synthesis content, playback duration, triggering mechanism, and device identification information. The server sends this data packet to the target point reading device through the communication channel established by the point reading device (such as Wi-Fi direct connection, Bluetooth link, or local area network communication protocol). After receiving this instruction, the point reading device automatically calls the local audio decoder to play the corresponding prompt voice to guide the user to know that a code value conflict has occurred and immediately go to the mobile phone to complete the book selection operation.

[0018] Step S5: After the user selects the target book on the mobile terminal, the server receives the selection result, binds the target book selected by the user to the point reading code, and stores it in the user configuration database; Specifically, after the user confirms the target book, the mobile terminal application constructs a selection result data packet containing multiple keyword fields. This data packet includes the book ID selected by the user currently, the session identifier previously issued by the server to uniquely identify the current interaction session, and the selection marker type indicating whether the user checks the "Set as default permanently". The data packet is sent back to the server through an encrypted channel. After receiving the data packet, the server starts the data verification module to perform field integrity, data type legality, and session identifier matching checks to ensure that the data has not been tampered with and is within the valid interaction cycle. If the verification passes, the valid user selection information in the data packet is extracted. The server calls the complete metadata information of the target book from the resource database according to the book ID specified in the valid user selection information, including the book title, publisher, version number, book category, cover image index, and the corresponding audio resource file path, etc., and constructs a standardized book resource mapping record based on this. This mapping record identifies the resource path of the target book and marks the context tags related to the user behavior in the current system. The server then calls the QR code association module to bind the resource mapping record with the QR code value that triggers this operation, and combines the user ID, QR code value, and target book ID to form a user-specific QR code mapping entry. This entry defines the association logic of "for a certain user, when using a certain code value, give priority to a certain book", forming the binding relationship between the QR code and the target book. Analyze whether the option of "Set as default permanently" is checked in the user selection behavior. If the user chooses to set this relationship as the default, the highest priority is given to this binding entry and the validity period is set to "permanent", thus generating a persistent mapping record; otherwise, only the temporary binding relationship in the current session is retained and the short-term cache expiration time is set. For the persistent mapping record, the server calls the database write interface to write this record into the QR code mapping table in the user configuration database, and uses the index optimization mechanism to ensure fast retrieval by the double primary keys of the user ID and QR code. At the same time, the context information such as the timestamp of this operation, the user device identifier, and the operation network environment (such as Wi-Fi / cellular network) is synchronously written into the user QR code history record table for subsequent behavior modeling and usage scenario analysis. After the above storage operation is completed, a confirmation response message is constructed, the content of which includes a binding success prompt and the book resource index selected this time, and the audio resource file corresponding to the target book is attached. The audio resource file is sent to the QR code reading device using a dedicated transmission protocol (such as fragmented HTTP or real-time WebSocket stream). After receiving the confirmation instruction sent by the server, the QR code reading device immediately switches to the target audio resource playback state, realizing the audio feedback output of the book content selected by the user, thus constructing an instant QR code reading feedback mechanism after the user operation, ensuring that the user perceives the system response and verifying whether the selected book meets the expectation, so as to form a complete interaction closed-loop and positive experience cycle on the user side.

[0019] Step S6: According to the binding relationship, when the user subsequently uses the same point-reading code for point-reading, the server automatically returns the audio resource file of the target book without the user having to select again.

[0020] Specifically, when the user uses the point-reading pen to point-read the same book again, the point-reading device generates a new point-reading request data packet based on the current point-reading behavior. This data packet contains key fields such as the point-reading code value, device identifier, user ID (or session identity passed by the login status), timestamp, and point-reading position, and is sent to the point-reading request processing module on the server side through the network. After receiving this subsequent point-reading request data packet, the server parses and processes it, extracts the point-reading code value contained in the current request, and constructs a query key in the form of a composite key value of "point-reading code value + user ID" in combination with the attached user identifier for precise matching in the user configuration database. When the system retrieves the mapping record corresponding to this composite key in the database, it calls the point-reading code mapping table, which records all the personalized binding relationships between each user and their devices for point-reading codes and books, including the association time, priority level, whether it is the default binding, binding source, and context information, etc. After successful matching, one or more book binding records associated with this point-reading code value are obtained. Since the user may have bindings from different sources for the same code value (for example, temporary bindings and default bindings coexist), the server performs a priority sorting operation on the records in the query result. The priority setting follows clear rules: the permanent default binding relationship takes precedence over the temporary binding, the one with a closer binding time takes precedence, and the binding source being the user's active selection takes precedence over the system's automatic inference, so as to select the binding record with the highest priority as the target book, and thus construct a resource location request, which specifies the book resource ID and the target resource type (here it is the audio resource), and is ready to initiate a data request to the resource server. After the resource location request is submitted, the server connects to the audio resource file server, obtains the corresponding audio file path from the distributed audio storage system based on the requested target book information, calls the audio scheduling module to load the audio file in slices into the buffer area, and performs processing such as encoding format matching, rate control, and data packaging on it to generate an audio data stream that conforms to the real-time transport protocol (such as RTP, WebRTC, or a custom long connection transport protocol). The server continuously pushes this audio data stream to the device side through the real-time communication channel established with the point-reading device. After receiving the stream data, the point-reading device decodes and starts playing, thus automatically completing book recognition and audio resource matching and backhaul without any additional operations by the user, significantly optimizing the user experience.

[0021] In the embodiments of the present invention, the server automatically detects the point-reading code conflict situation and simultaneously sends prompt messages to the point-reading device and the mobile phone terminal, realizing the timely discovery and active notification of conflicts, avoiding user doubts and troubles, and clearly guiding users to take subsequent operations. By using the user point-reading behavior data collected by the point-reading device, a user point-reading context relationship graph is constructed. By analyzing the user's point-reading habits and content preferences, the matching degree between the conflicting book and the current point-reading situation is intelligently calculated, providing personalized recommendation sorting for users, greatly improving the accuracy and convenience of selecting conflicting books. A collaborative working mechanism between the point-reading device and the mobile terminal is realized. By combining the voice prompt of the point-reading device and the visual interface of the mobile terminal, an intuitive and clear book selection process is created, reducing the user operation threshold and enhancing the interaction experience. The corresponding relationship of the user's book selection result is established and stored in the user configuration database, realizing the function of "select once, remember forever". When the user uses the same point-reading code later, the system automatically returns the audio resource of the selected book without repeating the selection process, greatly simplifying the user operation steps. By analyzing the user's historical selection behavior and point-reading habits, a dynamic weight mechanism is established for conflicting books, and a higher priority is given to the content types preferred by users, continuously optimizing the user's point-reading experience and making the system more in line with the user's personalized needs. The present invention completely realizes the solution of point-reading conflicts and the acquisition of audio resources through the network, avoiding the cumbersome operation of downloading audio resources through USB in the traditional solution, eliminating the dependence on personal computers, enabling all users to conveniently solve the point-reading conflict problem, and improving the universality and user-friendliness of the system.

[0022] In a specific embodiment, the process of executing step S1 may specifically include the following steps: The point-reading device captures the coordinate position, point-reading pressure value, and point-reading stay time of the user's point-reading pen tip through a built-in position sensor, generating original point-reading data; The original point-reading data is transmitted to the server, and the original point-reading data is sorted in time series to obtain structured point-reading time-series data; According to the structured point-reading time-series data, the position change law of adjacent point-reading operations is analyzed, and the user's point-reading direction and point-reading rhythm characteristics are extracted to generate a user point-reading behavior feature record.

[0023] Specifically, the point-reading device captures the coordinate position, point-reading pressure value, and point-reading dwell time of the user's point-reading pen tip through a built-in position sensor and converts them into digital signals with time stamps and spatial characteristics. To this end, the point-reading device integrates at least three types of sensing elements, including a high-resolution position detection sensor, a pressure sensor, and a high-precision timer. The point-reading pen uses infrared matrix scanning, capacitive array, or optical image recognition technology to achieve two-dimensional coordinate positioning of the pen tip on the paper surface. For each point-reading operation, at the moment the pen tip touches the paper surface, the device captures the X-axis and Y-axis values of this position and attaches a time stamp of the current operation. At the same time, through the pressure-sensitive sensor built into the pen body, the pressure value applied by the user during this point-reading is detected, and this value reflects the force exerted by the user during point-reading, assisting in inferring the user's attention intensity or mis-touch probability. To accurately judge the integrity and duration of the point-reading operation, the device measures the contact duration between the pen tip and the paper surface through a high-precision clock. The duration from when the pen tip is pressed down to when it is lifted constitutes the point-reading dwell time, and this indicator plays an important role in behavioral rhythm modeling. The data in the above three dimensions constitute an original point-reading data record, and the format includes: [time stamp, X coordinate, Y coordinate, pressure value, dwell time], which is bound to the device ID and the current session number for subsequent tracking and analysis on the server side. After the data collection is completed, the point-reading device uploads the collected original point-reading data to the cloud server in real time or in batches using a stable communication protocol. The data transmission channels include Wi-Fi, Bluetooth, or a dedicated low-power communication protocol (such as BLE), and the data transmission needs to support an encryption mechanism to protect user privacy. After the server receives the data, it sorts each original point-reading data record in ascending order according to the time stamp to form structured point-reading time-series data with time continuity. After obtaining the structured point-reading time-series data, the server calls a dedicated behavior analysis module to perform pairwise comparison and sliding window processing on adjacent point-reading records to analyze the point-reading trajectory trend and rhythm pattern of the user over a period of time. In terms of direction analysis, the direction vector between two consecutive point-reading coordinates is calculated, and the direction vector is represented by (X2 - X1, Y2 - Y1) and converted into an angle θ = arctangent((Y2 - Y1) / (X2 - X1)) for statistical analysis. When multiple direction vectors show consistency within a certain time window (for example, from the upper left to the lower right, or from top to bottom), the system can identify the point-reading direction trend of the user. In addition, by analyzing the change trend of the angle between consecutive direction vectors, it can be identified whether the user is performing jump point-reading, sequential reading, or back-and-forth review and other different reading paths. In terms of rhythm analysis, using the time stamp difference and dwell time data, the interval time and point-reading duration of each point-reading are calculated. If the user frequently and continuously points to different coordinate positions within a few seconds and the dwell time for each time is short, it is inferred that it is a quick browsing type of operation; if the user stays at a certain position for a long time and there is a regular interval between point-reading operations, it is inferred that it is a deep intensive reading type of behavior.Through the joint analysis of parameters such as the interval time, the duration of point reading, and the fluctuation of pressure values, multiple indicators reflecting the user's rhythm characteristics are extracted, including the average point-reading rhythm (number of point readings per minute), the maximum continuous stay time, the rhythm volatility, etc. The server combines the two types of features of the point-reading direction and the point-reading rhythm to construct a record of the user's current point-reading behavior characteristics. This record is represented in vector form and includes dimensions such as [main direction angle, direction stability coefficient, average rhythm frequency, rhythm volatility, average stay time, average pressure value, behavior type label], etc., and is stored in the user behavior model for various scenarios such as subsequent point-reading code conflict matching calculation, content recommendation, and behavior prediction. During long-term operation, this behavior model is continuously and dynamically updated, replacing old features with new data to form an adaptive description of the user's real usage habits, improving the system's response intelligence and processing accuracy in the face of complex situations such as point-reading conflicts.

[0024] Among them, according to the structured point-reading time series data, the position change law of adjacent point-reading operations is analyzed, and the user's point-reading direction and point-reading rhythm characteristics are extracted to form the user's point-reading behavior pattern, including: performing vector difference calculation on adjacent point-reading coordinates in the structured point-reading time series data to generate a point-reading displacement sequence, and dividing it into three types of features: horizontal displacement, vertical displacement, and mixed displacement according to the displacement direction; performing time window segmentation processing on the point-reading displacement sequence, and calculating the point-reading density, average displacement distance, and direction consistency index within each fixed time window to obtain a set of point-reading rhythm characteristic parameters; constructing a Markov state transition model based on the point-reading displacement sequence, representing the user's point-reading habit with the transition probability matrix of adjacent point-reading operations, and generating a point-reading direction pattern fingerprint; performing multi-feature fusion on the set of point-reading rhythm characteristic parameters and the point-reading direction pattern fingerprint, and using an adaptive weighted algorithm to assign dynamic weights to the features in different scenarios to construct a user's point-reading behavior feature identifier; applying hierarchical clustering analysis to the user's point-reading behavior feature identifier to identify the differences in the user's point-reading behavior at different times and different book types, forming a multi-mode point-reading behavior portrait; associating the multi-mode point-reading behavior portrait with context information such as time, environment, and book type, and establishing a conditional-triggered point-reading behavior prediction model to complete the construction of the user's point-reading behavior pattern.

[0025] In a specific embodiment, the process of executing step S2 may specifically include the following steps: The server receives the point-reading code sent by the point-reading device, and inputs the point-reading code into the point-reading code mapping database for query and matching to obtain a matching result; Count the number of corresponding books in the matching result, determine whether it exceeds a single corresponding threshold, determine whether there is a conflict situation where multiple books correspond to the same point-reading code, and generate a conflict detection result; Based on the conflict detection result, analyze the characteristic data of the conflicting books and perform conflict classification to obtain a conflict type determination result; Generate a formatted conflict message based on the conflict type determination result, push the formatted conflict message to the mobile terminal, embed a unique session identifier in the pushed message, and send a conflict prompt message to the mobile terminal.

[0026] Specifically, after the user's actual point reading behavior is completed, the point reading device sends a network request to the server. This request contains a structured data packet, including the currently triggered point reading code value, device ID, user identity identifier (if any), point reading timestamp, and auxiliary fields such as the attached point reading position coordinates. After receiving the request data packet, the receiving module of the server performs data integrity verification and field legality checks to ensure that the data has not been truncated or forged. After confirming that it is correct, it extracts the point reading code value from it as the main query parameter. The server inputs the extracted point reading code into the point reading code mapping database in the core of the system. This database is an efficient index structure established specifically for point reading resources. Each point reading code corresponds to multiple resource entries, including fields such as book ID, book name, publisher, version number, suitable reading age group, upload time, etc. The server performs a fast matching operation in the database through hash mapping or B+ tree indexing and returns all resource records that exactly match this point reading code, forming a set of matching result sets. Subsequently, the system calls the statistics module to count the quantity of this set. If the statistical result shows that the number of returned books is equal to 1, it means that this point reading code only corresponds to a single resource in the current database and there is no conflict. The server will directly return the corresponding book resource and record the request completion status. If the statistical result is greater than 1, it means that multiple books share the same instruction code, that is, a point reading code conflict occurs. At this time, the server generates a conflict detection result and marks the point reading code corresponding to this point reading behavior as a conflict status. After detecting the code value conflict, it enters the conflict analysis stage. The server extracts the metadata fields of all book entries in the matching results one by one and performs feature comparison and analysis on these fields. The main analysis contents include the content category of the book (such as: story, popular science, language learning, etc.), publishing unit, graphic structure, number of pages, suitable reading object age group, etc. By setting a series of standardized classification rules, the degree of difference between books in multiple dimensions is vectorized and modeled, and the similarity between each pair of book entries is calculated. The content and structure overlap degree is quantified using methods such as Euclidean distance, cosine similarity, or fuzzy logic matching. According to the analysis results, the conflict situation is divided into three types: the first is a low-level conflict, which means that the content types or target audiences of the books are significantly different and the misrecognition probability is low; the second is a medium-level conflict, which means that the book categories are similar but the page structures or styles are different, and there is still a certain degree of difficulty in judgment; the third is a high-level conflict, which means that the contents, covers, and styles of multiple books are extremely similar or even different versions from the same publisher, and it is very easy for users to make wrong selections. It is necessary to guide them to make a clear choice first. The final output of the conflict type is fed back to the message generation module in the form of a judgment result. After obtaining the conflict type judgment result, the system starts the formatted message construction module to generate a structured conflict prompt message.The message structure contains multiple fields. First, there is the text description of the conflict code value, the number of conflicting books, and the conflict level. Second, for each conflicting book, there are the book ID, the URL of the cover thumbnail, the book title, the publisher, the publication time, and the target audience label. All conflicting books are sorted in descending order according to the matching degree (based on historical page-turning behaviors) so that the most likely targets can be presented first during subsequent display. To ensure the uniqueness of the interactive session, the session management module is called to generate a random but unique session identifier, which will be bound to this page-turning behavior and used to track the status during subsequent user interaction confirmation or selection replacement. The formatted conflict message is packaged into a standardized data object (such as a JSON structure) and sent to the user's mobile App through the established WebSocket long connection between the server and the mobile terminal application or the mobile push platform channel (such as FCM, APNs). The push strategy uses an immediate reminder form to pop up a notification box or directly guides the user to the embedded selection interface module.

[0027] In a specific embodiment, the process of executing step S3 may specifically include the following steps: The server extracts the user's most recent page-turned book type, page-turning time pattern, page-turning location distribution, and historical selection data from the user's page-turning behavior feature record to construct the current page-turning context feature vector. Extract the book feature data for each conflicting book in the conflicting book list and construct the book feature vector for each conflicting book based on the book feature data. Calculate the similarity between the current page-turning context feature vector and the book feature vector of each conflicting book, and use the weighted Bayesian algorithm to assign different weights to each feature factor to obtain the initial matching score. Adjust according to the initial matching score in combination with the user's historical selection preference data to generate the target matching degree score, and sort the conflicting book list in descending order based on the target matching degree score to generate the sorted list of conflicting books.

[0028] Specifically, when the user triggers a point-reading behavior, the server calls the behavior records in the recent period from the user's point-reading behavior feature database. This record is structured to store the operation habits and content preferences demonstrated by the user in the past several point-reading sessions. The server extracts information from four key dimensions: First, the book type tags that the user has point-read, such as picture books, popular science, pinyin, English enlightenment, nursery rhymes, or story books, etc. These type tags are automatically labeled by the system during the book upload stage, or dynamically learned and adjusted during the user's operations; Second, the time distribution characteristics presented by the user's point-reading behavior, by counting the point-reading frequencies at different times of the day and on different dates of the week (weekdays and weekends) to form a point-reading time pattern vector; Third, the point-reading position distribution characteristics on the page. By counting the dense areas of multiple point-reading coordinates in the book page, it can be identified whether the user is inclined to the title area, the text area, the illustration area, or the corner area, etc., and normalize the modeling of these position hotspots; Fourth, the historical selection behavior of the user in past code value conflict situations, including the actual selections made by the user among multiple candidate books on multiple occasions. These selections are recorded in the database as "point-reading code - book ID - confirmation status" triples, which are used to form a historical selection behavior vector. After the above four types of information are unified and integrated, they are encoded into a current point-reading situation feature vector with a fixed dimension, such as a real number vector of length N, where each dimension corresponds to a type preference, a probability of a type of time interval, a spatial distribution weight, or a weight value of a historical preference label. Perform structured analysis and feature extraction operations on all candidate book entries that generate conflicts in this point-reading event. The server calls the book database and extracts the basic attributes and content features of each book, including the type tags to which the book belongs (which can be multi-label intersections), the publishing unit, the content structure description (such as the picture-text ratio, the average number of words per page, whether there is picture and audio synchronization), the suitable reading age group, the language category, and other fields, and converts these attributes into a book feature vector in a unified format based on a preset label vector system. Calculate the similarity between the current point-reading situation feature vector and the book feature vector of each conflicting book. Since the information in different dimensions has different importance to the result, the weighted Bayesian algorithm is used to assign different weights to each dimension. Initialize a set of Bayesian prior weight parameters, where the user type preference weight is set to 0.25, the time pattern matching weight is set to 0.15, the spatial position weight is set to 0.30, and the historical selection weight is set to 0.30, and the sum is 1, indicating that these four types of characteristic factors jointly determine the user's selection preference. Calculate the relative proximity between the current point-reading vector value and the book feature vector value for each dimension, and use this proximity as the likelihood value for this dimension. Calculate the posterior probability according to Bayes' theorem, that is, the probability that the user selects a certain candidate book under the current user behavior background, and finally output a set of initial matching scores, where each conflicting book corresponds to a value between 0 and 1, and the larger this value is, the more it matches the user's current behavior.Based on the initial matching scores, user historical selection preferences are introduced for adjustment and optimization. The system retrieves the user's historical selection records related to the current point - reading code value from the user configuration database. If a certain book has been selected by the user multiple times and consistently in the same or similar conflict situations, the system will increase the initial matching score of this book by weighting, such as increasing it by 10% - 20%. At the same time, a punitive reduction operation is implemented for books that have never been selected or whose selection has been revoked. If the user has not formed a selection record under this code value before, the system makes an inference and correction by referring to the preference tags formed in similar books, similar types, or similar time periods, and obtains the target matching degree scores of each conflicting book in the current behavior context. All conflicting books are sorted in descending order according to the target matching degree scores, and a structured sorted list of conflicting books is generated. This list indicates the matching scores of each book and also includes auxiliary display data such as sorting serial numbers, book IDs, cover thumbnail links, publication information, and suitable - for - reading tags. This sorted list will be synchronously pushed to the display interface of the user's mobile terminal, and when displayed on the client side, the book with the highest score is defaultly placed at the top of the interface.

[0029] Among them, the similarity between the current point - reading situation feature vector and the book feature vector of each conflicting book is calculated. The weighted Bayesian algorithm is used to assign different weights to each feature factor to obtain the initial matching score, including: dimension - normalizing the current point - reading situation feature vector and the book feature vector of the conflicting book respectively, unifying the features with different dimensions into the standard unit space to obtain the standardized feature vectors; constructing a feature importance evaluation matrix, calculating the discrimination ability of each feature dimension through the information gain rate, assigning an initial weight coefficient to each feature dimension to generate a feature weight vector; implementing a weighted Bayesian conditional probability model based on the feature weight vector, calculating the posterior probability of the user selecting each conflicting book given the current point - reading situation, and forming a book matching probability distribution; introducing the user's historical selection data as prior knowledge, converting the historical selection frequency into a prior probability distribution, and adjusting the matching probability value by combining with the posterior probability through Bayes' formula; applying a non - linear transformation function to the matching probability value to make the probability difference more significant, and setting an adaptive threshold parameter for different conflict types to generate a normalized matching score; adjusting the timeliness of the normalized matching score by combining with a time - decay factor, making the influence weight of the most recent point - reading behavior on the matching result higher than that of early behaviors, and outputting the initial matching score.

[0030] In a specific embodiment, the process of executing step S4 may specifically include the following steps: The server sends the sorted list of conflicting books to the mobile terminal, and based on the sorted list of conflicting books, places the recommended book in a prominent position at the top of the interface, and arranges the other conflicting books in descending order of matching degree in the lower area to generate a user - interface display template; Send the user interface display template to the mobile terminal application, compress each book image in the user interface display template, and add interactive elements to form an interactive selection interface; Build a point reading device control data packet containing voice synthesis instructions based on the interactive selection interface and generate a voice playback control instruction; Send the voice playback control instruction to the point reading device through the communication channel of the point reading device, trigger the point reading device to play voice prompt information, and guide the user to select books on the mobile terminal.

[0031] Specifically, the server sends the conflict book sorted list to the mobile terminal and starts the interface layout construction module. This module is led by the visual logic of the user interface, extracts the book item with the highest recommendation priority from the sorting result as the top recommended book, and places this book in the top area of the user interface in a highlighted style. This highlighted style includes a large-size cover image display, an independent display block, a prominent color block border, recommended label annotation, and operation guide text, etc. The remaining conflict books are arranged in descending order of the matching degree generated by the server and are successively arranged below the recommendation block, forming a vertically scrollable list structure that conforms to the attention guidance principle. This interface structure is constructed with a double-layer logic of "recommended book area + candidate book area", binds a unique book ID and image resource index to each book, and at the same time binds logic event response hooks for interface interaction, such as click to select, set as default, expand detailed information, etc. The system generates a user interface display template according to this structured layout template, and its content is encapsulated in the form of JSON or XML, and includes book cover links, publisher information, suitable reading labels, matching scores, recommendation identifiers, image priority fields, etc. The server calls the image compression and rendering optimization module to perform image compression processing on all book cover images to be displayed on the mobile terminal interface. This process adopts JPEG compression, WebP format conversion or image multi-resolution adaptation strategy to balance image display clarity and mobile terminal transmission performance, and controls the image size within a reasonable range (such as not exceeding 150KB) on the premise of ensuring the recognition of core content. At the same time, combined with the current resolution and device memory status of the mobile terminal, a suitable image size version is selected. After the compression processing, a unified interactive response component is added to each cover image, such as HTML button binding, sliding response area, selection mark, trigger identifier, etc., to ensure that subsequent logic calls can be triggered after the user clicks the image, such as confirming the selected book, setting it as the default book or viewing detailed information. At this time, the interactive selection interface has completed resource integration and structure generation on the server side. The server transmits this interactive display template to the mobile terminal application of the bound user through the WebSocket real-time channel or the HTTP push interface. After receiving the template, the mobile terminal restores the template content to a visual interface by the front-end rendering engine (such as React Native, Flutter or native UI module), and adapts it in combination with factors such as device resolution, user language environment, system theme color, etc., to ensure that the finally presented interface has clear logic, quick response, and prominent visual focus. At the same time, to achieve the auditory assistance for user behavior guidance, the server synchronously generates a voice prompt for the reading device to play. The server calls the built-in text-to-speech (TTS) synthesis module to convert the structured prompt text into natural voice waveform data. The voice data adopts the PCM, MP3 or AAC format, and controls parameters such as speech rate, intonation, and pause rhythm to conform to the auditory acceptance habit.After the voice file is synthesized, it is encapsulated into a device control data packet. The data packet contains fields such as the target device ID, voice file path or content, play trigger timing, play mode (whether to loop), play priority, etc., which constitute the voice play control instruction. The server sends the voice play control instruction to the target reading pen device through the communication channel established between the reading pen device and the server (which can be a Wi-Fi direct connection, a local area network UDP broadcast, a Bluetooth Low Energy BLE link, or a private protocol channel). After receiving the instruction, the reading pen device parses the instruction content through the local voice decoding module and starts the audio play process, controlling the built-in speaker to play the prompt voice, guiding the user to know that there is a conflict in the current reading pen behavior, and prompting the user to go to the mobile phone to select the corresponding book content. If the user does not complete the selection within a certain period of time, the system is set to repeat the play or send a reminder prompt to ensure that the guiding process is not interrupted.

[0032] In a specific embodiment, the process of executing step S5 may specifically include the following steps: Receive the selection result data packet sent by the mobile terminal, which contains the selected book ID, session identifier, and selection marker type, and perform data verification on the selection result data packet to obtain valid user selection information; Extract the detailed information of the selected target book according to the valid user selection information and generate a book resource mapping record; Associate the book resource mapping record with the corresponding reading pen code to construct a user-specific reading pen code mapping entry, forming a binding relationship between the reading pen code and the target book; Check whether the user checks the option of permanently setting as the default for the binding relationship. If checked, set the binding priority to the highest level and generate a persistent mapping record; Write the persistent mapping record into the reading pen code mapping table of the user configuration database, and at the same time update the user's reading pen history record table, recording the selection timestamp and selection environment information to complete the storage operation; Send a confirmation message and the audio resource file of the selected target book to the reading pen device, so that the reading pen device immediately plays the audio content of the user-selected target book, establishing an instant reading pen feedback mechanism.

[0033] Specifically, after the user completes the operation of selecting a conflicting book on the mobile terminal, the client encapsulates the selection result into a structured data packet and sends it to the server through the HTTP POST or WebSocket channel. The data packet includes three key fields: one is the unique identifier ID of the selected book (such as Book_ID), the second is the unique session identifier (Session_ID) in the current interaction process, and the third is the type of operation mark attached by the user to this selection behavior (such as whether to set it as the default, whether it is only valid for this time, etc.). This mark is represented in the form of a boolean value or an enumeration. After the receiving module of the server receives the data packet, it executes a data verification program to check the existence, format legality, and semantic consistency of each field. For example, the system will verify whether Book_ID belongs to the legal options in the current conflict list, whether Session_ID is consistent with the previously pushed session identifier, and whether the selection mark type conforms to the logical specification. After all fields pass the verification, the system marks it as valid selection data and uses it as the basis for the current interaction action. After obtaining the valid user selection information, the server calls the resource mapping service module to extract the complete metadata of the target book from the main book database according to Book_ID, including fields such as book title, publisher, publication time, language version, book content path (i.e., audio file path), suitable reading age range, and illustration resource address, and summarizes and encapsulates this information into a book resource mapping record. Call the user configuration management module to correspond and bind the current point reading code that triggered this selection to this book resource mapping record, and construct a "point reading code - book ID" mapping entry exclusive to the current user. The mapping entry structure contains fields such as user identifier, point reading code value, book ID, binding timestamp, binding source (user selection, system recommendation, etc.), binding status, binding validity period, and priority weight, which are used to represent the book content selection corresponding to the current user under a specific code value. After the binding relationship is established, it is judged whether the user has checked the option of "set as default" during the operation. If the system detects that the mark type is permanently effective, the priority of this binding relationship is set to the highest level, and at the same time the binding validity period is set to be long-term effective (such as permanently or in years), and the system marks this binding relationship as a "persistent mapping record"; if the user does not set it as the default, the system sets this binding as a temporary mapping, with a lower priority and a validity period set to the current session or within 24 hours. The system generates the final binding record according to the above rules. Write this binding relationship into the point reading code mapping table in the user configuration database, with User_ID and Code_Value as the combined primary key, to ensure that the default selection of the user under this code value can be quickly located during query.Meanwhile, the point-reading behavior history record module is called to add a new behavior record in the point-reading behavior log table, recording the timestamp of this binding operation, the selected book ID, whether it is set as the default, the triggering method (such as mobile phone selection), the interaction session identifier, the operation terminal type (iOS / Android), and the network status during the operation (Wi-Fi / 4G / offline cache), so as to perform user behavior modeling and exception handling analysis later. After completing the persistent storage of all binding relationships and behavior records, the server constructs a point-reading feedback control data packet, which consists of two parts. One is the confirmation message, sending the flag information of "binding successful, please play the book audio" to the point-reading device. The other is the audio resource data, using the streaming audio transmission format (such as WebSocket real-time stream, HTTP segmented download, RTP packet stream), and pushing the audio file of the target book (usually in chapter segment format) to the point-reading device in real time as needed. After receiving the feedback data packet, the point-reading device parses the confirmation message and confirms the successful binding status, and then calls the local player module to load and play the audio data segment pushed by the server, realizing the instant voice output of the book content corresponding to the user's current point-reading behavior. This output not only meets the user's expectations but also establishes a "selection-response-verification" closed-loop interaction mechanism between the point-reading device and the server.

[0034] In a specific embodiment, the process of executing step S6 may specifically include the following steps: The server receives the subsequent point-reading request data packet sent by the point-reading device according to the binding relationship, parses the subsequent point-reading request data packet, and extracts the point-reading code value to be processed; Combines the point-reading code value to be processed with the user identifier to form a query key value, and performs a matching search in the user configuration database to obtain the point-reading code mapping query result; Performs a priority sorting on the point-reading code mapping query result, selects the target book record with the highest priority, and generates a resource location request; Accesses the audio resource file server according to the resource location request, obtains the audio resource file of the corresponding target book, generates an audio data stream, and sends the audio data stream to the point-reading device through the real-time transport protocol without the user having to select again.

[0035] Specifically, when the user uses the pen reader to perform a point reading operation on a book with a previous code value conflict again after completing the initial binding, the software system inside the point reading device generates a subsequent point reading request data packet based on the current user's pen tip contact event. The structure of this data packet contains multiple key fields, among which the most core one is the currently collected point reading code value. In addition, it also includes the device identifier, user identity identifier, point reading trigger timestamp, point reading coordinates (for auxiliary determination), and the network status identifier of the current device (such as connectivity, transmission protocol version, etc.). The point reading device sends this data packet to the receiving module of the cloud server in real time through a preset data channel. After receiving the data packet, the server routes it to the point reading behavior parsing service component. In the behavior parsing stage, the server parses the content of the data packet into a structured format, extracts the point reading code value as the query core for this resource search, and extracts the user identifier for subsequent configuration mapping positioning. Combine the point reading code value and the user identifier to generate a combined query key value, that is, the (User_ID, Code_Value) binary tuple. This key value is used to accurately retrieve the point reading binding relationship established by the current user under this code value from the user configuration database. Use this combined key as the primary key to call the point reading code mapping table and perform a conditional query operation. All user-level personalized mapping relationships are recorded in this table, and each record contains fields such as user ID, point reading code, book ID, establishment time, binding source, priority weight, validity period, and status flag. After the retrieval operation is completed, the server obtains all the binding records related to this user and this code value. Since the user may have established multiple bindings (such as different versions of the book or the coexistence of temporary and default), the server executes a priority sorting logic on the result set. The priority sorting rules are defined as follows: The binding set as "permanent default" has the highest priority, followed by "temporary default", then "recent temporary binding", and then "system inferred association" or "historical operation rollback". At the same time, the system introduces dimensions such as binding creation time and usage frequency as auxiliary factors to further refine the priority, and finally selects the book record with the highest priority as the target book corresponding to the current point reading code value. After the server determines this target book record, it constructs a resource location request. The structure of this request contains fields such as book ID, resource type (audio), chapter structure identifier (such as whether to load the whole book or page by page), request source (i.e., the point reading device identifier), encryption identifier (for verifying data access rights), etc. This request will be submitted to the audio resource file server, and the resource management platform will perform resource path parsing and file loading. During this process, the resource platform accesses the corresponding resource storage path according to the book ID and retrieves the corresponding audio file as needed. The audio file is either the total audio file of the whole book or the segmented audio divided by chapters, using standard audio coding formats such as AAC or MP3, and is positioned in combination with page numbers or anchor information. The server loads the audio file into the buffer cache area and calls the audio encapsulation module to generate a data stream format that conforms to the real-time transmission standard.Specific transmission protocols adopt WebSocket real-time audio push, HTTP segmented download, or packetized stream transmission mode based on RTP (Real-Time Protocol) according to device capabilities and network conditions, and perform sequence control, lost packet retransmission control, and bandwidth adjustment control on data packets during the transmission process to ensure that the point-reading device can efficiently and stably receive audio data in different network environments. After receiving the audio data stream, the point-reading device automatically calls the local audio playback engine to decode and play the audio. Without any further confirmation or selection by the user, the correct audio content corresponding to the page being point-read can be heard. During this process, the internal state of the device will also be synchronously updated, such as recording the playback state, current playback resource, remaining playback time, whether it is interrupted, etc., for subsequent behavior statistics and personalized optimization of the system.

[0036] In this embodiment, the process of the server calculating the matching degree of each conflicting book with the current point-reading situation based on the record of user point-reading behavior characteristics further includes an optimization method for selecting conflicting books based on multi-agent reinforcement learning, including: setting each conflicting book as an independent decision-making agent, creating an agent state space, including book type characteristics, content structure characteristics, user historical selection frequency, and current point-reading context characteristics, to construct a multi-agent decision-making environment; assigning a dynamic reward function to each book agent, constructing a composite reward mechanism based on user selection behavior, point-reading satisfaction feedback, and system operation efficiency indicators to guide the agent to make an optimal selection decision; designing and implementing a social attribute perception mechanism to enable each book agent to perceive the content correlation, user preference similarity, and point-reading situation matching degree between it and other book agents, forming a relationship perception tensor; based on the relationship perception tensor, constructing an attention-weighted multi-head adaptive network to dynamically integrate the relationship information between book agents and generate a context-enhanced agent state representation; for the context-enhanced agent state representation, applying a policy gradient algorithm to train the decision-making policy network of each book agent to optimize the selection probability distribution of the agent in different point-reading situations; using an experience replay buffer pool to store historical interaction data, sampling the data according to time-decaying weights, and regularly updating the value evaluation network and policy network parameters of each book agent; introducing a decentralized asynchronous advantage actor-critic framework to enable each book agent to learn in parallel while maintaining global coordination and accelerating the policy convergence speed; designing an agent cooperation and competition balance mechanism to achieve collaborative optimization of the selection between conflicting book agents by adjusting the weight ratio of the global value function and the local value function; applying a curriculum learning strategy to gradually transition from a simple point-reading scenario to a complex conflict situation, enabling the agent system to gradually master point-reading conflict resolution strategies at different difficulty levels; in each actual process of selecting conflicting books, activating the most suitable agent strategy based on the current point-reading situation, calculating the final conflicting book matching score, and generating an optimal recommendation result with theoretical guarantee.

[0037] The method for selecting conflict e-books in the embodiments of the present invention has been described above. Next, the device for selecting conflict e-books in the embodiments of the present invention will be described. Please refer to Figure 2 , an embodiment of the device for selecting conflict e-books in the embodiments of the present invention includes: An acquisition module, configured to acquire the time and position data of the user's e-book reading operation by the e-book reading device and transmit the data to the server, and generate a record of the user's e-book reading behavior characteristics; A judgment module, configured to receive the e-book reading code sent by the e-book reading device by the server and judge whether there is a conflict situation where multiple e-books correspond to the same e-book reading code. When there is a conflict, a conflict prompt message is sent to the mobile terminal; A calculation module, configured to calculate the matching degree between each conflict e-book and the current e-book reading scenario according to the record of the user's e-book reading behavior characteristics by the server, and generate a sorted list of conflict e-books; A guidance module, configured to send the sorted list of conflict e-books to the mobile terminal by the server, and at the same time send a voice prompt message to the e-book reading device to guide the user to select an e-book on the mobile terminal; An establishment module, configured to receive the selection result by the server after the user selects a target e-book on the mobile terminal, establish a binding relationship between the target e-book selected by the user and the e-book reading code, and store the relationship in the user configuration database; An automatic return module, configured to automatically return the audio resource file of the target e-book according to the binding relationship by the server when the user subsequently reads an e-book using the same e-book reading code, without the user having to select again.

[0038] Through the collaborative cooperation of the above-mentioned various components, the server automatically detects the point-reading code conflict situation and simultaneously sends prompt messages to the point-reading device and the mobile phone terminal, realizing the timely discovery and active notification of conflicts, avoiding user confusion and annoyance, and clearly guiding users to take subsequent operations. By using the user point-reading behavior data collected by the point-reading device, a user point-reading context relationship graph is constructed. By analyzing the user's point-reading habits and content preferences, the matching degree between the conflicting book and the current point-reading situation is intelligently calculated, providing personalized recommendation sorting for users, and greatly improving the accuracy and convenience of selecting conflicting books. The collaborative working mechanism between the point-reading device and the mobile terminal is realized. By combining the voice prompt of the point-reading device and the visual interface of the mobile terminal, an intuitive and clear book selection process is created, reducing the user operation threshold and enhancing the interaction experience. The corresponding relationship of the user's book selection result is established and stored in the user configuration database to realize the function of "select once, remember permanently". When the user uses the same point-reading code later, the system automatically returns the audio resources of the selected book without repeating the selection process, greatly simplifying the user operation steps. By analyzing the user's historical selection behavior and point-reading habits, a dynamic weight mechanism is established for conflicting books, and a higher priority is given to the content types preferred by users, continuously optimizing the user's point-reading experience and making the system more in line with the user's personalized needs. The present invention completely realizes the solution of point-reading conflicts and the acquisition of audio resources through the network, avoiding the cumbersome operation of downloading audio resources through USB in the traditional solution, eliminating the dependence on personal computers, enabling all users to conveniently solve the point-reading conflict problem, and improving the universality and user-friendliness of the system.

[0039] Referring to Figure 3 , an embodiment of the present invention further provides a computer device, which may be a server, and its internal structure may be as Figure 3 shown. The computer device includes a processor, a memory, a display screen, an input device, a network interface, and a database connected through a system bus. Among them, the processor of the computer device is used 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 the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store the corresponding data in this embodiment. The network interface of the computer device is used to communicate with external terminals through a network. When the computer program is executed by the processor, the above method is implemented.

[0040] Those skilled in the art can understand that Figure 3 the structure shown in

[0041] An embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the above method is implemented. It can be understood that the computer-readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0042] Those of ordinary skill in the art can understand that all or part of the processes in the above-described method embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the above-described method embodiments. Among them, any reference to a memory, storage, database, or other medium provided by the present invention and used in the embodiments can include non-volatile and / or volatile memories. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or an external cache. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM, etc.

[0043] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described system, system, and unit can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0044] When the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0045] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for selecting conflict books for point reading, characterized in that, Including: The point-reading device collects the time and position data of the user's point-reading operation and transmits it to the server to generate a record of the user's point-reading behavior characteristics; The server receives the point-reading code sent by the point-reading device and determines whether there is a conflict situation where multiple books correspond to the same point-reading code. When there is a conflict, it sends a conflict prompt message to the mobile terminal; The server calculates the matching degree of each conflicting book with the current point-reading context based on the record of the user's point-reading behavior characteristics, and generates a sorted list of conflicting books; The server sends the sorted list of conflicting books to the mobile terminal, and at the same time sends a voice prompt message to the point-reading device to guide the user to select a book on the mobile terminal; After the user selects the target book on the mobile terminal, the server receives the selection result, establishes a binding relationship between the target book selected by the user and the point-reading code, and stores it in the user configuration database; Based on the binding relationship, when the user uses the same point-reading code for point-reading subsequently, the server automatically returns the audio resource file of the target book without the user having to select again.

2. The method for selecting conflicting point-reading books according to claim 1, wherein The point-reading device collects the time and position data of the user's point-reading operation and transmits it to the server to generate a record of the user's point-reading behavior characteristics, including: The point-reading device captures the coordinate position, point-reading pressure value, and point-reading stay time of the user's point-reading pen tip through a built-in position sensor to generate raw point-reading data; Transmit the raw point-reading data to the server, and perform time series sorting on the raw point-reading data to obtain structured point-reading time series data; According to the structured point-reading time series data, analyze the position change law of adjacent point-reading operations, extract the user's point-reading direction and point-reading rhythm characteristics, and generate a record of the user's point-reading behavior characteristics.

3. The method for selecting conflicting point-reading books according to claim 1, characterized in that, The server receives the point-reading code sent by the point-reading device and determines whether there is a conflict situation where multiple books correspond to the same point-reading code. When there is a conflict, it sends a conflict prompt message to the mobile terminal, including: The server receives the point-reading code sent by the point-reading device, and inputs the point-reading code into the point-reading code mapping database for query and matching to obtain a matching result; Count the number of corresponding books in the matching result, determine whether it exceeds the single corresponding threshold, determine whether there is a conflict situation where multiple books correspond to the same point-reading code, and generate a conflict detection result; Analyze the characteristic data of the conflicting books based on the conflict detection result and perform conflict classification to obtain a conflict type determination result; Generate a formatted conflict message according to the conflict type determination result, and push the formatted conflict message to the mobile terminal, embed a unique session identifier in the pushed message, and send a conflict prompt message to the mobile terminal.

4. The method for selecting conflicting point-reading books according to claim 1, wherein The server calculates the matching degree of each conflicting book with the current point-reading context based on the record of the user's point-reading behavior characteristics, and generates a sorted list of conflicting books, including: The server extracts the user's most recent point-read book type, point-reading time pattern, point-reading position distribution, and historical selection data from the record of the user's point-reading behavior characteristics to construct a current point-reading context feature vector; Extract the book characteristic data for each conflicting book in the list of conflicting books, and construct a book characteristic vector for each conflicting book based on the book characteristic data; Calculate the similarity between the current point-reading context feature vector and the book feature vectors of each conflicting book, and use the weighted Bayesian algorithm to assign different weights to each feature factor to obtain an initial matching score; Adjust according to the initial matching score in combination with the user's historical selection preference data to generate a target matching degree score, and sort the conflicting book list in descending order based on the target matching degree score to generate a sorted list of conflicting books.

5. The method for selecting conflicting point-reading books according to claim 1, characterized in that, The server sends the sorted list of conflicting books to the mobile terminal, and at the same time sends a voice prompt message to the point-reading device to guide the user to select a book on the mobile terminal, including: The server sends the sorted list of conflicting books to the mobile terminal, and based on the sorted list of conflicting books, places the recommended book at a prominent position at the top of the interface, and arranges the other conflicting books in the lower area according to the matching degree to generate a user interface display template; Send the user interface display template to the mobile terminal application, and perform compression processing on each book image in the user interface display template and add interactive elements to form an interactive selection interface; Construct a point-reading device control data packet containing a voice synthesis instruction based on the interactive selection interface and generate a voice playback control instruction; Send the voice playback control instruction to the point-reading device through the communication channel of the point-reading device to trigger the point-reading device to play a voice prompt message to guide the user to select a book on the mobile terminal.

6. The method for selecting conflicting point-reading books according to claim 1, wherein After the user selects a target book on the mobile terminal, the server receives the selection result, and establishes a binding relationship between the target book selected by the user and the point-reading code and stores it in the user configuration database, including: Receive a selection result data packet sent by the mobile terminal containing the selected book ID, session identifier, and selection marker type, and perform data verification on the selection result data packet to obtain valid user selection information; Extract the detailed information of the selected target book according to the valid user selection information to generate a book resource mapping record; Associate the book resource mapping record with the corresponding point-reading code to construct a user-specific point-reading code mapping entry to form a binding relationship between the point-reading code and the target book; Check whether the user has checked the option of permanently setting as the default for the binding relationship. If checked, set the binding priority to the highest level to generate a persistent mapping record; Write the persistent mapping record into the point-reading code mapping table of the user configuration database, and at the same time update the user point-reading history record table to record the selection timestamp and selection environment information to complete the storage operation; Send a confirmation message and the audio resource file of the selected target book to the point-reading device to enable the point-reading device to immediately play the audio content of the target book selected by the user and establish an instant point-reading feedback mechanism.

7. The method for selecting conflicting point-reading books according to claim 1, wherein According to the binding relationship, when the user uses the same point-reading code for point-reading subsequently, the server automatically returns the audio resource file of the target book without the user having to select again, including: The server receives a subsequent point-reading request data packet sent by the point-reading device according to the binding relationship, parses the subsequent point-reading request data packet, and extracts the point-reading code value to be processed; Combine the to-be-processed dot-reading code value with the user identification to form a query key value, perform a matching search in the user configuration database, and obtain a dot-reading code mapping query result; Perform a priority sorting on the dot-reading code mapping query result, select the target book record with the highest priority, and generate a resource location request; Access the audio resource file server according to the resource location request, obtain the audio resource file corresponding to the target book, generate an audio data stream, and send the audio data stream to the dot-reading device through the Real-Time Transport Protocol without the user having to select again.

8. A device for selecting conflicting point-reading books, characterized in that, For executing the method for selecting dot-reading conflict books as described in any one of claims 1-7, the device for selecting dot-reading conflict books includes: An acquisition module, configured to acquire the time and position data of the user's dot-reading operation by the dot-reading device and transmit them to the server, generating a user dot-reading behavior feature record; A judgment module, configured to receive the dot-reading code sent by the dot-reading device by the server and judge whether there is a conflict situation where multiple books correspond to the same dot-reading code. When there is a conflict, send a conflict prompt message to the mobile terminal; A calculation module, configured to calculate the matching degree of each conflict book with the current dot-reading situation according to the user dot-reading behavior feature record by the server, generating a conflict book sorting list; A guidance module, configured to send the conflict book sorting list to the mobile terminal by the server, and at the same time send a voice prompt message to the dot-reading device to guide the user to select a book on the mobile terminal; An establishment module, configured to, after the user selects a target book on the mobile terminal, the server receives the selection result, establish a binding relationship between the target book selected by the user and the dot-reading code, and store it in the user configuration database; An automatic return module, configured to, according to the binding relationship, when the user uses the same dot-reading code for dot-reading subsequently, automatically return the audio resource file of the target book by the server without the user having to select again.

9. A computer device, characterized in that, Comprising a memory and a processor, the memory stores a computer program that can run on the processor, and when the processor executes the computer program, it implements the method for selecting dot-reading conflict books as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, Having a computer program stored thereon, and when the computer program is run by the processor, the processor is caused to execute the method for selecting dot-reading conflict books as described in any one of claims 1 to 7.

Images