Commodity information publicizing method, system and device based on two-dimensional code and medium

Abstract

This application provides a method, system, device, and medium for displaying product information based on QR codes. The method includes: acquiring structured information data stored in the QR code image on the product packaging, wherein the structured information data includes key product information entities and meta-descriptive data associated with these key information entities; analyzing the structured information data based on multi-dimensional context parameters obtained locally from the user's device to generate a personalized strategy vector; dynamically selecting and serializing information entities in the structured information data according to the personalized strategy vector to generate an interactive information set; and generating multimodal interaction commands based on the interactive information set. This method can dynamically adjust the information display format according to the user's real-time status and needs, improving the equality of information access and interactive experience for special groups in various environments.

Description

Technical Field

[0001] This invention belongs to the field of information processing technology, and in particular relates to a method, system, device and medium for displaying product information based on QR codes. Background Technology

[0002] With the widespread adoption of IoT and smart terminal technologies, QR codes have become a key carrier connecting physical goods with digital information. Against this backdrop, technologies have emerged that utilize QR codes to extend the display of product information. The common model involves scanning the code to obtain a network address pointing to a cloud server, which then loads a dynamic webpage containing product details, promotional content, or traceability information. This approach places the complex information rendering and interaction logic on the server side, providing a relatively rich visual page. Currently, access to basic product information such as production date and expiration date relies on this online query model or is entirely dependent on physical printing on the packaging.

[0003] However, current online access methods have inherent limitations: First, they heavily rely on stable, high-speed network connections, becoming completely ineffective in environments with poor signal or no network, thus losing information accessibility. Second, the webpage content delivered by servers is typically designed to meet the general browsing needs of ordinary users; its information structure is complex and its layout varied, lacking a stable and user-friendly structured design for the specific needs of visually impaired individuals or the elderly, and it cannot be adapted for real-time, personalized interaction on the local device. Meanwhile, some alternative solutions aimed at achieving offline access, such as directly encoding plain text information into QR codes, while avoiding network dependence, present information in an extremely rigid and simplistic way, merely mechanically transferring information from printed materials to the screen. They cannot dynamically respond or adjust according to the user's real-time interactive intentions or physical condition, essentially remaining a one-way, static data broadcast. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, system, device, and medium for displaying product information based on QR codes to address the aforementioned technical issues. This method can proactively adapt to user characteristics and respond to multimodal information flows that reflect real-time interactive intentions, thereby truly achieving barrier-free access to product information that is equal, inclusive, and user-controllable.

[0005] Firstly, this application provides a method for disclosing product information based on QR codes, including:

[0006] Obtain the structured information data stored in the QR code image on the product packaging, wherein the structured information data includes key information entities of the product and meta-description data associated with the key information entities of the product;

[0007] Based on multi-dimensional context parameters obtained locally from the user device, structured information data is analyzed to generate personalized strategy vectors. The multi-dimensional context parameters include at least one of user feature parameters, device capability parameters, and real-time environment parameters obtained from the user device.

[0008] Based on personalized strategy vectors, information entities in structured information data are dynamically selected and serialized to generate interactive information sets;

[0009] Multimodal interaction instructions are generated based on interactive information sets, wherein the multimodal interaction instructions are used to instruct the user device to generate and output visual and / or auditory information from structured information data.

[0010] In one embodiment, structured information data is analyzed based on multi-dimensional context parameters obtained locally from the user device to generate a personalized policy vector, including:

[0011] The enabled status of the screen reader or magnifier function is used as the first user characteristic parameter, and the user's historical interaction preference data is obtained by accessing the user's local configuration file as the second user characteristic parameter.

[0012] Read the user device's screen physical size and resolution as the first device capability parameter, and enumerate the text-to-speech engines available locally on the user device and their supported speech attributes as the second device capability parameter.

[0013] The ambient light sensor of the user equipment is called to obtain the ambient illuminance value as the first real-time environmental parameter, and the ambient audio of a preset duration is collected through the microphone of the user equipment and its average sound pressure level is calculated as the second real-time environmental parameter.

[0014] The first user feature parameter, the second user feature parameter, the first device capability parameter, the second device capability parameter, the first real-time environment parameter, and the second real-time environment parameter are normalized and vectorized, and the processed parameter vectors are concatenated to obtain the context feature vector.

[0015] Information structure features are extracted from the meta-description layer of structured information data, and the context feature vector is concatenated with the information structure features to obtain the concatenated feature vector.

[0016] The concatenated feature vector is input into a pre-defined decision model for forward computation, and the decision model outputs a personalized strategy vector.

[0017] In one embodiment, the concatenated feature vector is input into a pre-defined decision model for forward computation, and the decision model outputs a personalized policy vector, including:

[0018] The decision model is a fully connected neural network, and the number of neurons in the input layer of the decision model is consistent with the dimension of the concatenated feature vector;

[0019] The concatenated feature vector is input into the input layer of the decision model, and then nonlinearly transformed through at least one hidden layer of the decision model. Finally, a multidimensional policy vector is output at the output layer of the decision model.

[0020] The first dimension of the strategy vector is used to characterize the dominant information output mode, the second dimension is used to characterize the level of detail in the information broadcast, the third dimension is used to index the predefined broadcast order template, and the fourth dimension is used to control the activation status of the interaction channel.

[0021] In one embodiment, information entities in structured information data are dynamically selected and serialized according to a personalized strategy vector to generate an interactive information set, including:

[0022] Based on the third dimension value in the personalized strategy vector, the corresponding broadcast order template identifier is found in the meta-description layer of the structured information data;

[0023] Based on the broadcast order template identifier, the semantic identifier sequence of the information entity to be processed is obtained from the order list defined in the meta-description layer;

[0024] Following the order determined by the semantic identifier sequence, information entity data objects corresponding to each semantic identifier are extracted sequentially from the information entity layer of the structured information data.

[0025] For each extracted information entity data object, based on the level of detail represented by the second dimension value in the personalized strategy vector, text content that matches the level of detail is selected as the core display content from the text detail mapping associated with the information entity data object.

[0026] Query whether the current information entity data object is registered as an interactive hotspot in the meta description layer. If so, generate an interactive context object based on the semantic identifier of the information entity data object and the predefined list of interactive actions in the meta description layer, and associate and bind the interactive context object with the core display content to form an information unit.

[0027] All information units obtained through sequential processing are combined to generate an interactive information set.

[0028] In one embodiment, generating multimodal interaction instructions based on an interactive information set includes:

[0029] Based on the arrangement order of each information entity in the interactive information set and its core display content, connect words based on natural language generation templates are inserted to generate a coherent voice broadcast text script.

[0030] Based on the dimension values ​​in the personalized policy vector used to represent the dominant output modality and the real-time environment parameters in the multi-dimensional context parameters, the target speech rate is calculated using the following formula:

[0031]

[0032] in, To achieve the target speaking speed, As the system's baseline speech rate, This is the environmental factor adjustment coefficient. The ambient noise sound pressure level is obtained from real-time environmental parameters. For reference noise sound pressure level, For user fitness adjustment coefficient, The sharpness compensation factor is calculated based on user historical interaction preference data;

[0033] Generate multimodal interactive instructions to drive the local text-to-speech engine. These instructions include a text script for voice playback and a target speech rate. The multimodal interactive instructions are used to instruct the user device to synthesize and play audio.

[0034] Secondly, this application also provides a QR code-based product information disclosure system, including:

[0035] The data acquisition module is used to acquire the structured information data stored in the QR code graphic on the product packaging. The structured information data includes key product information entities and meta-description data associated with the key product information entities.

[0036] The personalized analysis module is used to analyze structured information data based on multi-dimensional context parameters obtained locally from the user device and generate personalized strategy vectors. The multi-dimensional context parameters include at least one of user feature parameters, device capability parameters and real-time environment parameters obtained from the user device.

[0037] The interaction strategy module is used to dynamically select and serialize information entities in structured information data according to personalized strategy vectors, and generate interactive information sets.

[0038] The information disclosure module is used to generate multimodal interaction instructions based on the interactive information set. The multimodal interaction instructions are used to instruct the user device to generate and output visual and / or auditory information from structured information data.

[0039] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the above-described method for displaying product information based on QR codes.

[0040] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the above-described method for displaying product information based on QR codes.

[0041] The aforementioned QR code-based product information display method, system, device, and medium dynamically generate personalized strategy vectors by parsing the structured information data in the product QR code and combining it with multi-dimensional context parameters of the user's device (including user characteristics, device capabilities, and real-time environment). Based on this, information entities are intelligently selected and serialized, ultimately transforming into multimodal interaction commands. This technical solution achieves adaptive presentation of product information without network dependence, and can dynamically adjust the information display format according to the user's real-time status and needs, improving the equality of information access and interactive experience for special groups in various environments. Attached Figure Description

[0042] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 A flowchart illustrating a product information display method based on QR codes, provided in an embodiment of the present invention;

[0044] Figure 2 This is a schematic diagram of a QR code-based product information disclosure system provided in an embodiment of the present invention. Detailed Implementation

[0045] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0046] First, a brief introduction to the terms used in the embodiments of this application will be given.

[0047] In this solution, structured information data specifically refers to a collection of digital information with a predetermined organizational format decoded from the QR code graphic on product packaging. It differs from traditional QR codes that only contain simple text or a single web link. Its core lies in including not only the specific numerical or textual information of key product attributes (such as production date, shelf life, and ingredients)—the "information entities"—but also "meta-descriptive data" describing how these entities are organized, presented, and interacted with. This structural design endows the data with self-interpretability and programmable processing capabilities, providing the necessary data foundation for subsequent personalized adaptation.

[0048] Multi-dimensional context parameters refer to a collection of quantitative or state information acquired locally and in real-time from the user device executing this method, representing various influencing factors in the current information access scenario. These parameters are not from a single source but encompass parameters reflecting the user's own characteristics (such as accessibility usage status), parameters reflecting the device's hardware and software capabilities (such as screen specifications and speech synthesis support), and parameters reflecting the external physical environment (such as ambient light and noise levels). The fusion of these parameters constitutes the basis for the system's perception of the "current context" and is a prerequisite for achieving dynamic adaptation of information output.

[0049] Multimodal interaction commands are a set of instructions or data that can directly drive user device hardware or system services to perform specific actions. They are generated based on a set of interactive information and explicitly instruct the device how to transform the processed information into a form perceptible to the user. For example, a command might include a complete transcript and speech rate parameters for synthesized speech, or text content and its layout coordinates for enlarged display on the screen. Essentially, these commands act as a bridge connecting information processing logic and physical output devices, ensuring that personalized, interactive intentions are accurately translated into a real audiovisual experience.

[0050] Based on the above definitions, the implementation environment of a QR code-based product information display method provided in this application is described. Indicatively, the implementation environment includes: a sensor, a terminal, a standard pressure generator, and a processor. The sensor may include a light sensor, a sound sensor, a distance sensor, a gravity sensor, or a positioning sensor, etc.; the processor may be a central processing unit, an artificial intelligence chip, or a multi-core processor, and is not limited thereto.

[0051] Based on the above definitions and implementation environment, the application scenarios of the embodiments of this application are described. The product information display method based on QR codes provided in the embodiments of this application can be applied to scenarios including but not limited to the following:

[0052] In fields such as pharmaceuticals and health products, where the accuracy and accessibility of information are extremely important, the application of this technical solution is even more significant. When using medications, elderly people or patients often face the risk of misreading information due to small font sizes and obscure terminology. This method, by scanning the QR code on the medicine box, not only provides enlarged and enhanced instructions for use and contraindications, but also allows for confirmation through segmented, slow-speed voice prompts. If users have questions about a particular piece of information, they can click on the corresponding area of ​​the screen to trigger a more accessible explanation. This localized, interactive information delivery method is independent of network conditions, providing an additional layer of digital protection for safe medication use.

[0053] In the field of public facilities, this technology is applicable to accessible tour guide systems. When visually impaired individuals scan the QR codes of museum exhibits, the technology can transform the structured data of the artifacts into layered audio narration based on the connection status of the guide cane (device capability), the ambient noise level (environmental parameters), and preset information density preferences (user characteristics). In noisy environments, it automatically condenses core information, while in quiet environments, it outputs detailed background descriptions. This solution frees cultural facilities from dependence on fixed audio guide equipment, achieving a truly inclusive and autonomous visiting experience.

[0054] This is merely an illustrative example; the product information disclosure method based on QR codes provided in this application embodiment can also be applied to other application scenarios. This is only an example and does not limit the specific application scenarios.

[0055] In one exemplary embodiment, such as Figure 1 As shown, a method for displaying product information based on QR codes is provided. This embodiment illustrates the application of this method to a terminal in the aforementioned implementation environment. It is understood that this method can also be applied to a server, or to a system including both a terminal and a server, and is implemented through the interaction between the terminal and the server. The method includes the following steps 101 to 104:

[0056] Step 101: Obtain the structured information data stored in the QR code graphic on the product packaging. The structured information data includes key product information entities and meta-description data associated with the key product information entities.

[0057] Specifically, the user device's image acquisition module scans the QR code on the product packaging, then parses and obtains the pre-stored structured information data within the QR code. This structured information data is organized using a preset standardized data format. The key product information entities cover core content such as the product's production date, ingredients, specifications, and instructions for use. Meta-description data corresponds to the type, meaning, data priority, and association rules of each key information entity, ensuring that the data is identifiable and operable. This structured information data can be obtained by directly decoding it using the QR code parsing algorithm pre-installed on the device, without relying on a cloud server. Even in a network-free environment, data extraction can be completed quickly, effectively overcoming the limitations of traditional online queries that depend on the network.

[0058] Step 102: Based on the multi-dimensional context parameters obtained locally from the user device, analyze the structured information data to generate a personalized strategy vector, wherein the multi-dimensional context parameters include at least one of user feature parameters, device capability parameters and real-time environment parameters obtained from the user device.

[0059] Specifically, this method performs in-depth analysis of structured information data based on multi-dimensional context parameters retrieved locally from the user's device, thereby generating a personalized strategy vector. User characteristic parameters can be obtained through device system authorization, including user-preset information presentation preferences, accessibility settings (such as visual impairment adaptation mode and elderly mode), and historical interaction records. Device capability parameters are obtained through device hardware interface detection, covering hardware indicators such as display resolution, audio output power, and local computing performance, as well as supported software interaction functions. Real-time environmental parameters are collected in real-time by the device's built-in sensors, including dynamic information such as ambient light intensity, noise level, and network signal strength. During the analysis, this method employs a local lightweight algorithm to extract features and assign weights to the multi-dimensional context parameters, transforming the core features of the user, device, and environment into quantifiable strategy factors that the device can recognize. The resulting personalized strategy vector clarifies the core direction of information processing, providing a precise decision-making basis for subsequent personalized adaptation.

[0060] Step 103: Based on the personalized strategy vector, dynamically select and serialize the information entities in the structured information data to generate an interactive information set.

[0061] Specifically, based on the generated personalized strategy vector, information entities in structured information data are dynamically selected and serialized to generate an interactive information set. The dynamic selection process is based on the priority weights in the strategy vector to filter out the core information that best matches the current user needs, device capabilities, and environmental conditions, while eliminating redundant and irrelevant content. The serialization and arrangement determine the order in which information is presented according to the user's information acquisition habits, logical connections between information, and environmental adaptation requirements. In terms of implementation, the selection and sorting of information entities can be automatically completed through a preset information filtering rule base and arrangement algorithm, combined with the parameter instructions in the personalized strategy vector. The resulting interactive information set not only fits the current usage scenario but also supports subsequent user interactions such as clicks and swipes, breaking the limitations of traditional static information transmission.

[0062] Step 104: Generate multimodal interaction instructions based on the interactive information set, wherein the multimodal interaction instructions are used to instruct the user device to generate and output visual and / or auditory information from structured information data.

[0063] Specifically, this method generates multimodal interaction instructions based on the generated interactive information set. These instructions specify the specific modal type, parameter settings, and interaction logic of the information presentation according to the requirements of the personalized strategy vector. The visual presentation instructions can specify the font size, color contrast, and graphic display format of the text, while the auditory presentation instructions can specify the speech rate, volume, and language type of the voice broadcast. It also supports the separate or combined output of visual and auditory modalities. The multimodal interaction instructions are converted into executable operation signals by the device's local instruction parsing module, instructing the user device to output structured information data in a way that adapts to the current scenario, ensuring that the user can obtain information in the most convenient way. For example, it automatically switches to high-contrast large font display in bright light environments and increases the volume of voice broadcast and reduces the speech rate in noisy environments.

[0064] The aforementioned QR code-based product information display method dynamically generates personalized strategy vectors by parsing the structured information data in the product's QR code and combining it with multi-dimensional context parameters of the user's device (including user characteristics, device capabilities, and real-time environment). Based on this, it intelligently selects and serializes information entities, ultimately transforming them into multimodal interaction commands. This technical solution achieves adaptive presentation of product information without network dependence, and can dynamically adjust the information display format according to the user's real-time status and needs, improving the equality of information access and interactive experience for special groups in various environments.

[0065] In one embodiment, structured information data is analyzed based on multi-dimensional context parameters obtained locally from the user device to generate a personalized policy vector, including:

[0066] The enabled status of the screen reader or magnifier function is obtained as the first user characteristic parameter, and the user's historical interaction preference data is obtained by accessing the user's device's local configuration file as the second user characteristic parameter.

[0067] Specifically, this method calls the accessibility application programming interface (API) provided by the user's device operating system to query whether high-contrast visual aids such as screen readers or magnifiers are activated. Their activation status is used as the first user characteristic parameter to infer potential user needs. Simultaneously, the method accesses an encrypted configuration file in the device's local storage. This file anonymously records the user's interaction preferences throughout their application usage, such as the average selected speech rate, frequently skipped information types, or frequently triggered "repeat listening" actions. This statistical data is extracted as the second user characteristic parameter. By combining real-time status and historical behavior, this method constructs a comprehensive understanding of user needs, rather than relying on single-dimensional guesswork.

[0068] The screen physical size and resolution of the user device are read as the first device capability parameter, and the locally available text-to-speech engines and their supported speech attributes are enumerated as the second device capability parameter.

[0069] Specifically, this method performs a detailed evaluation of the user device's hardware and software capabilities as objective constraints for the adapted output. For example, the method reads the physical size, pixel density, and current resolution of the display system; these data together constitute the first device capability parameter that determines the layout of visual information and the limits of font scaling. Simultaneously, the method enumerates all text-to-speech engines installed locally on the device and obtains a list of their supported speech attributes, such as available languages, synthesized timbre, and whether adjustable speech rate and tone are supported. This information serves as the second device capability parameter, ensuring that the generated voice commands can be effectively executed on the current device.

[0070] The ambient light sensor of the user device is used to obtain the ambient illuminance value as the first real-time environmental parameter, and the ambient audio of a preset duration is collected through the microphone of the user device and its average sound pressure level is calculated as the second real-time environmental parameter.

[0071] Specifically, the method calls the ambient light sensor to obtain the current ambient illuminance value as the first real-time environmental parameter for evaluating screen visibility. At the same time, after user authorization, the method briefly activates the device's microphone to collect an ambient audio sample lasting several seconds, and then calculates the average sound pressure level of the sample as the second real-time environmental parameter for quantifying the ambient noise level. This real-time data enables the information output strategy to dynamically respond to changes in the external environment, such as enhancing display contrast under strong light or automatically increasing voice volume in noisy environments.

[0072] The first user feature parameter, the second user feature parameter, the first device capability parameter, the second device capability parameter, the first real-time environment parameter, and the second real-time environment parameter are normalized and vectorized, and the processed parameter vectors are concatenated to obtain the context feature vector.

[0073] Specifically, various parameters are normalized to eliminate the influence of dimensions and unify the numerical range. For example, Boolean state values ​​are mapped to 0 or 1, and continuous values ​​such as screen size are transformed to a specific range through minimum-maximum scaling or Z-score normalization. Each normalized parameter is converted into a fixed-dimensional numerical vector, and all parameter vectors are concatenated in a predetermined order to obtain a comprehensive, high-dimensional context feature vector. This vector is a digital condensation of the current "human-machine-environment" overall situation.

[0074] Information structure features are extracted from the meta-description layer of structured information data, and the context feature vector is concatenated with the information structure features to obtain the concatenated feature vector.

[0075] Specifically, in order to incorporate the structural characteristics of the product information itself into decision-making, this method also extracts features from the meta-description layer of the decoded structured information data, namely, information structure features. These features may include summary information such as the total number of information entities, the distribution ratio of entities in each category, and whether there are interactive hotspot definitions. This method then concatenates the aforementioned context feature vector with this information structure feature vector to form a more comprehensive concatenated feature vector, which simultaneously contains a joint representation of the external context and the internal data structure.

[0076] The concatenated feature vector is input into a pre-defined decision model for forward computation, and the decision model outputs a personalized strategy vector.

[0077] For example, this method inputs the concatenated feature vector into a pre-built decision model on the device for forward computation. The decision model adopts a lightweight neural network architecture, and the model parameters obtained through offline training and optimization are pre-built on the device, without relying on cloud computing power. During the model computation, the concatenated feature vector is transformed and weighted through a fully connected layer, and the features are non-linearly mapped by an activation function. The output is a personalized strategy vector that includes core contents such as information filtering rules, presentation mode selection, and parameter configuration scheme. This vector can directly guide the dynamic processing and interactive output of subsequent information, and the model forward computation process is responsive, ensuring that users can obtain personalized information services without long waiting times.

[0078] In one embodiment, the concatenated feature vector is input into a pre-defined decision model for forward computation, and the decision model outputs a personalized policy vector, including:

[0079] The decision model is a fully connected neural network, and the number of neurons in the input layer of the decision model is consistent with the dimension of the concatenated feature vector.

[0080] Specifically, the decision model used in this method is a lightweight fully connected neural network. After offline training, the model parameters are pre-installed on the user's device, eliminating the need for cloud computing power and enabling fast local inference. The number of neurons in the input layer of the decision model is strictly consistent with the dimension of the concatenated feature vector, ensuring that every feature information in the concatenated feature vector can be accurately received by the input layer neurons, avoiding feature loss or calculation errors due to dimension mismatch. At the same time, this fully connected neural network reduces device power consumption and storage usage while ensuring computational accuracy by simplifying the network structure and optimizing the selection of activation functions.

[0081] The concatenated feature vector is input into the input layer of the decision model, and then undergoes a nonlinear transformation through at least one hidden layer of the decision model. Finally, a multidimensional policy vector is output at the output layer of the decision model.

[0082] For example, this method inputs concatenated feature vectors into the input layer of a decision model. The input layer neurons sequentially pass feature information to at least one hidden layer. The hidden layer performs nonlinear transformations on the feature vectors using predefined activation functions, achieving deep extraction and fusion of complex features. For instance, the ReLU activation function is used to enhance the model's nonlinear fitting ability, and batch normalization is used to stabilize gradient propagation during training, improving the effectiveness of feature transformation. After multiple rounds of feature processing in the hidden layers, a multidimensional policy vector is output at the output layer of the decision model. Each dimension of this vector corresponds to a specific information presentation control rule, providing precise guidance for the personalized output of subsequent information.

[0083] The first dimension of the strategy vector is used to characterize the dominant information output mode, the second dimension is used to characterize the level of detail in the information broadcast, the third dimension is used to index the predefined broadcast order template, and the fourth dimension is used to control the activation status of the interaction channel.

[0084] Specifically, in the strategy vector output by this method, the first dimension value explicitly represents the dominant information output modality through quantitative values. Different values ​​correspond to different modes such as visual presentation, auditory presentation, or audiovisual combination presentation, ensuring that the information presentation method is adapted to user needs and environmental conditions. The second dimension value represents the level of detail of information broadcasting through hierarchical values. The higher the value, the more comprehensive the information presented; the lower the value, the only core key information is retained, meeting the user's need for information conciseness in different scenarios. For example, if three levels are preset (brief, standard, and detailed), this dimension can use the output of three neurons of the Softmax activation function, selecting the neuron index with the highest probability as the level of detail. The third dimension value associates multiple broadcasting order templates pre-stored locally on the device through index values. Different templates correspond to different arrangement logics of information entities, realizing personalized adaptation of information presentation order. The fourth dimension value controls the activation status of the interaction channel through Boolean values, determining whether to support users to further interact with the information through clicks, swipes, and other operations, breaking the limitations of traditional static information transmission.

[0085] In one embodiment, information entities in structured information data are dynamically selected and serialized according to a personalized strategy vector to generate an interactive information set, including:

[0086] Based on the third dimension value in the personalized strategy vector, the corresponding broadcast order template identifier is searched in the meta-description layer of the structured information data.

[0087] Specifically, the third dimension value is extracted from the personalized strategy vector. This value serves as the index identifier for the broadcast order template, forming a one-to-one correspondence with the template identifiers pre-stored in the structured information data element description layer. This method traverses the template mapping table of the element description layer to accurately find the corresponding broadcast order template identifier based on the index identifier. The template identifier contains the arrangement logic rules of information entities, ensuring that the subsequent information presentation order matches the comprehensive needs of user characteristics, device capabilities, and the real-time environment, avoiding the low information acquisition efficiency problem caused by traditional fixed orders.

[0088] Based on the broadcast order template identifier, the semantic identifier sequence of the information entity to be processed is obtained from the order list defined in the meta-description layer.

[0089] Specifically, after obtaining the broadcast sequence template identifier, the method extracts a sequence of semantic identifiers for the information entities to be processed from a predefined sequence list in the meta-description layer. This sequence list is pre-constructed based on the logical relationships of product information, user acquisition habits, and scenario adaptation requirements. The semantic identifiers are standardized codes used to uniquely identify each information entity. During the extraction process, the method uses a semantic matching algorithm to confirm the completeness of the identifier sequence, ensuring no omissions or redundancies. This sequence clearly defines the presentation order of the information entities, providing a clear basis for the subsequent orderly extraction of information entity data.

[0090] Following the order determined by the semantic identifier sequence, information entity data objects corresponding to each semantic identifier are extracted sequentially from the information entity layer of the structured information data.

[0091] Specifically, this method traverses the information entity layer of structured information data according to the order determined by the semantic identifier sequence, using a structured data indexing mechanism. For each semantic identifier, a key-value pair matching method is used to accurately extract the corresponding information entity data object from the information entity layer. The data object contains the complete original data and associated attributes of the information entity. The extraction process employs a local, efficient retrieval algorithm, eliminating the need for cloud-based data supplementation and ensuring rapid data extraction even in offline environments. Furthermore, the extraction results strictly correspond to the identifier sequence, guaranteeing the accuracy of the information arrangement.

[0092] For each extracted information entity data object, based on the level of detail represented by the second dimension value in the personalized strategy vector, text content that matches the level of detail is selected as the core display content from the text detail mapping associated with the information entity data object.

[0093] Specifically, for each extracted information entity data object, this method first parses the second dimension value in the personalized strategy vector. This value represents the level of detail in the information broadcast, with different levels corresponding to different granularities of text display content. The method queries a pre-defined text detail mapping table within the information entity data object. This table pre-stores the association between different levels of detail and their corresponding text content. Based on the current level of detail, the matching text content is selected as the core display content, satisfying users' need for concise information while avoiding information overload or the loss of key information.

[0094] Query whether the current information entity data object is registered as an interactive hotspot in the meta description layer. If so, generate an interactive context object based on the semantic identifier of the information entity data object and the predefined list of interactive actions in the meta description layer, and associate and bind the interactive context object with the core display content to form an information unit.

[0095] Specifically, after determining the core display content, this method queries the interaction registration status of the current information entity data object in the meta-description layer to determine whether it is marked as an interaction hotspot. An interaction hotspot refers to a key information entity that the user may need to perform further operations to obtain more related information. If it is an interaction hotspot, the method calls a predefined list of interactive actions in the meta-description layer based on the semantic identifier of the information entity data object. The action list includes various interaction forms such as clicking to view details, voice inquiry, and jumping to related information. This generates an interaction context object containing action instructions and related data indexes. Then, the interaction context object is associated and bound to the core display content through a data binding algorithm, forming an information unit that combines display functions and interactive capabilities.

[0096] All information units obtained through sequential processing are combined to generate an interactive information set.

[0097] For example, this method sequentially combines all information units obtained by processing them according to the semantic identifier sequence, maintaining the order of each information unit consistent with the semantic identifier sequence during the combination process. Simultaneously, through data format standardization, it ensures the structural uniformity and compatibility of each information unit, generating an interactive information set. This information set not only contains core display content adapted to the current scenario but also integrates interactive functions, supporting users to perform further operations according to their needs, breaking the limitations of traditional static information sets.

[0098] In one embodiment, generating multimodal interaction instructions based on an interactive information set includes:

[0099] Based on the arrangement order of each information entity in the interactive information set and its core display content, connector words based on natural language generation templates are inserted to generate a coherent voice broadcast text script.

[0100] Specifically, this method analyzes the arrangement order of each information entity in the interactive information set, clarifies the logical relationship between information units, and extracts the core display content corresponding to each information entity to ensure the integrity and accuracy of the text material. The method calls the device's local pre-built natural language generation template library, which contains a set of connectors adapted to different information logical relationships. Based on the semantic relationship type of the information entities (such as causality, progression, parallelism, etc.), the method automatically selects appropriate connectors and inserts them between each core display content. Through a natural language concatenation algorithm, the discrete core display content and connectors are combined to generate a fluent and logically coherent voice broadcast text script, avoiding the problems of fragmented and abrupt information breaks in traditional voice broadcasts and improving the user's auditory reception experience.

[0101] Based on the dimension values ​​in the personalized policy vector used to represent the dominant output modality and the real-time environment parameters in the multi-dimensional context parameters, the target speech rate is calculated using the following formula:

[0102]

[0103] in, To achieve the target speaking speed, As the system's baseline speech rate, This is the environmental factor adjustment coefficient. The ambient noise sound pressure level is obtained from real-time environmental parameters. For reference noise sound pressure level, For user fitness adjustment coefficient, This is a clarity compensation factor calculated based on user historical interaction preference data.

[0104] Specifically, after generating the speech broadcast text script, this method extracts the dimension values ​​representing the dominant output modality from the personalized strategy vector to confirm that the current information output is speech as the core modality. Simultaneously, it retrieves real-time environmental parameters from the previously acquired multi-dimensional context parameters to clarify the key environmental data used for calculation. In this formula, This represents the system's preset baseline speaking speed, which is a default value suitable for quiet environments. It is a pre-calibrated environmental factor adjustment coefficient used to control the sensitivity of speech rate to the influence of environmental noise; It is a real-time measurement of the ambient noise sound pressure level obtained from real-time environmental parameters. It is a predefined reference noise sound pressure level, which serves as the reference point for adjustment; It is another calibrated user fitness adjustment coefficient; This is a clarity compensation factor calculated based on user historical interaction preference data stored locally on the device. The calculation of this factor is based on the frequency and density of "repeat" commands triggered in past conversations; the higher the frequency, the smaller the factor value, indicating that the user needs a slower speaking speed. For example, if the current environment is noisy and the user's history shows that they frequently need to listen repeatedly, the formula will automatically derive a clarity compensation factor lower than the baseline speaking speed. The value is designed to improve the intelligibility of the first broadcast by slowing down the speech to combat noise interference and cater to user habits.

[0105] Generate multimodal interactive instructions to drive the local text-to-speech engine. These instructions include a text script for voice playback and a target speech rate. The multimodal interactive instructions are used to instruct the user device to synthesize and play audio.

[0106] Specifically, this method generates multimodal interactive instructions to drive the local text-to-speech engine of the user device based on the generated speech playback text script and the calculated target speech rate. The instructions also include auxiliary parameters such as the volume reference value and language type of the speech playback. These parameters are all determined based on multi-dimensional context parameters. The multimodal interactive instructions are sent to the text-to-speech engine through the device's local instruction transmission interface, instructing the engine to synthesize an audio signal according to the script content, target speech rate and related parameters, and control the device's audio output module to play the audio. At the same time, it supports users to adjust the speech rate, pause or replay through preset interactive actions, so as to realize personalized and interactive output of voice information.

[0107] In summary, the product information display method based on QR codes provided in this application enables offline access to information by embedding structured information data containing key product information entities and related meta-descriptive data within the QR code, breaking the limitations of traditional network dependence. Furthermore, it extracts multi-dimensional contextual parameters from the user device, including user characteristics, device capabilities, and real-time environment, and after normalization, vectorization, and fusion with information structure features, inputs them into a locally pre-built lightweight decision model to generate personalized strategy vectors, clarifying core rules such as information output modality, level of detail, broadcast order, and interaction activation status. Based on these strategy vectors, through template matching, ordered extraction, detail adaptation, and interaction hotspot binding, it generates an interactive information set that fits the scenario requirements. Combining natural language generation technology and multi-factor weighted calculation, it generates multimodal interactive commands containing coherent speech scripts and adapted speech rates, driving the device to output information in visual and / or auditory forms. The entire solution utilizes a technical chain of local parsing of structured data, precise perception of multi-dimensional context, intelligent generation of personalized strategies, dynamic arrangement of interactive information, and multimodal adaptation output to construct a multimodal information flow that proactively adapts to user characteristics and responds to real-time interactive intentions. This effectively breaks through the limitations of static and generalized traditional information disclosure, truly achieving barrier-free access to product information that is equal, inclusive, and controllable by users.

[0108] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0109] Based on the same inventive concept, this application also provides a QR code-based product information disclosure system 10 for implementing the QR code-based product information disclosure method described above. The solution provided by this system is similar to the implementation scheme described in the above method. Therefore, the specific limitations of one or more embodiments of the QR code-based product information disclosure system 10 provided below can be found in the limitations of the QR code-based product information disclosure method described above, and will not be repeated here.

[0110] In one exemplary embodiment, such as Figure 2 As shown, a QR code-based product information disclosure system 10 is provided, including:

[0111] The data acquisition module 11 is used to acquire the structured information data stored in the QR code graphic on the product packaging, wherein the structured information data includes key information entities of the product and meta-description data associated with the key information entities of the product.

[0112] The personalized analysis module 12 is used to analyze structured information data based on multi-dimensional context parameters obtained locally from the user device and generate a personalized strategy vector. The multi-dimensional context parameters include at least one of user feature parameters, device capability parameters and real-time environment parameters obtained from the user device.

[0113] The interaction strategy module 13 is used to dynamically select and serialize information entities in structured information data according to personalized strategy vectors to generate an interactive information set.

[0114] The information disclosure module 14 is used to generate multimodal interaction instructions based on the interactive information set. The multimodal interaction instructions are used to instruct the user device to generate and output visual and / or auditory information from structured information data.

[0115] In one embodiment, the personalized analysis module 12 includes:

[0116] The user feature parameter acquisition unit is used to obtain the enabled status of the screen reader or magnifier function as the first user feature parameter, and to obtain the user's historical interaction preference data as the second user feature parameter by accessing the local configuration file of the user's device.

[0117] The device parameter acquisition unit is used to read the screen physical size and resolution of the user device as the first device capability parameter, and enumerate the text-to-speech engines available locally on the user device and their supported voice attributes as the second device capability parameter.

[0118] The environmental parameter acquisition unit is used to call the ambient light sensor of the user equipment to acquire the ambient illuminance value as the first real-time environmental parameter, and to collect ambient audio of a preset duration through the microphone of the user equipment and calculate its average sound pressure level as the second real-time environmental parameter.

[0119] The feature processing unit is used to normalize and vectorize the first user feature parameters, the second user feature parameters, the first device capability parameters, the second device capability parameters, the first real-time environment parameters, and the second real-time environment parameters, and then concatenate the processed parameter vectors to obtain the context feature vector.

[0120] The feature concatenation unit is used to extract information structure features from the meta-description layer of structured information data, and concatenate the context feature vector with the information structure features to obtain the concatenated feature vector.

[0121] The strategy generation unit is used to input the concatenated feature vector into a pre-set decision model for forward computation, and output a personalized strategy vector through the decision model.

[0122] In one embodiment, the policy generation unit is configured to perform the following steps:

[0123] The decision model is a fully connected neural network, and the number of neurons in the input layer of the decision model is consistent with the dimension of the concatenated feature vector;

[0124] The concatenated feature vector is input into the input layer of the decision model, and then nonlinearly transformed through at least one hidden layer of the decision model. Finally, a multidimensional policy vector is output at the output layer of the decision model.

[0125] The first dimension of the strategy vector is used to characterize the dominant information output mode, the second dimension is used to characterize the level of detail in the information broadcast, the third dimension is used to index the predefined broadcast order template, and the fourth dimension is used to control the activation status of the interaction channel.

[0126] In one embodiment, the interaction strategy module 13 includes:

[0127] The template matching unit is used to find the corresponding broadcast order template identifier in the meta-description layer of structured information data based on the third dimension value in the personalized strategy vector.

[0128] The sequence acquisition unit is used to obtain the sequence of semantic identifiers of the information entities to be processed from the sequence list defined in the meta-description layer based on the broadcast order template identifier;

[0129] The entity extraction unit is used to extract information entity data objects corresponding to each semantic identifier from the information entity layer of the structured information data in the order determined by the sequence of semantic identifiers.

[0130] The content filtering unit is used to select text content that matches the level of detail from the detailed text mapping associated with the information entity data object as the core display content for each extracted information entity data object, based on the level of detail represented by the second dimension value in the personalized strategy vector.

[0131] The interactive binding unit is used to query whether the current information entity data object has been registered as an interactive hotspot in the meta description layer. If so, an interactive context object is generated based on the semantic identifier of the information entity data object and the predefined list of interactive actions in the meta description layer. The interactive context object is then associated and bound with the core display content to form an information unit.

[0132] The set generation unit is used to combine all the information units obtained by sequential processing to generate an interactive information set.

[0133] In one embodiment, the information disclosure module 14 includes:

[0134] The script generation unit is used to insert connectors based on natural language generation templates according to the arrangement order of each information entity in the interactive information set and its core display content, and combine them to generate a coherent voice broadcast text script.

[0135] The speech rate calculation unit calculates the target speech rate based on the dimension values ​​representing the dominant output modality in the personalized policy vector and the real-time environment parameters in the multi-dimensional context parameters, using the following formula:

[0136]

[0137] in, To achieve the target speaking speed, As the system's baseline speech rate, This is the environmental factor adjustment coefficient. The ambient noise sound pressure level is obtained from real-time environmental parameters. For reference noise sound pressure level, For user fitness adjustment coefficient, The sharpness compensation factor is calculated based on user historical interaction preference data;

[0138] The instruction generation unit is used to generate multimodal interactive instructions that drive the local text-to-speech engine. The multimodal interactive instructions include a text script for voice playback and a target speech rate. The multimodal interactive instructions are used to instruct the user device to synthesize and play audio.

[0139] In one embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the QR code-based product information display method as described above.

[0140] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0141] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The components described as separate parts may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this disclosure according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0142] The above-described embodiments are merely illustrative of several implementation methods of the embodiments of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the embodiments of this application, and these modifications and improvements all fall within the protection scope of the embodiments of this application.

Claims

1. A method for displaying product information based on QR codes, characterized in that, The method includes: Obtain structured information data stored in the QR code image on the product packaging, wherein the structured information data includes key product information entities and meta-description data associated with the key product information entities; Based on multi-dimensional context parameters obtained locally from the user device, the structured information data is analyzed to generate a personalized strategy vector, wherein the multi-dimensional context parameters include at least one of user feature parameters, device capability parameters, and real-time environment parameters obtained from the user device. Based on the personalized strategy vector, information entities in the structured information data are dynamically selected and serialized to generate an interactive information set; Multimodal interaction instructions are generated based on the interactive information set, wherein the multimodal interaction instructions are used to instruct the user equipment to generate and output visual and / or auditory information from the structured information data.

2. The method according to claim 1, characterized in that, The step of analyzing the structured information data based on multi-dimensional context parameters obtained locally from the user device to generate a personalized policy vector includes: The activation status of the screen reader or magnifier function is obtained as the first user characteristic parameter, and the user's historical interaction preference data is obtained by accessing the local configuration file of the user device as the second user characteristic parameter. The screen physical size and resolution of the user device are read as the first device capability parameter, and the locally available text-to-speech engines and their supported speech attributes are enumerated as the second device capability parameter. The ambient light sensor of the user equipment is called to obtain the ambient illuminance value as the first real-time environmental parameter, and the ambient audio of a preset duration is collected through the microphone of the user equipment and its average sound pressure level is calculated as the second real-time environmental parameter. The first user feature parameter, the second user feature parameter, the first device capability parameter, the second device capability parameter, the first real-time environment parameter, and the second real-time environment parameter are normalized and vectorized, and the processed parameter vectors are concatenated to obtain the context feature vector. Information structure features are extracted from the meta-description layer of the structured information data, and the context feature vector is concatenated with the information structure features to obtain a concatenated feature vector. The concatenated feature vector is input into a preset decision model for forward computation, and the personalized strategy vector is output through the decision model.

3. The method according to claim 2, characterized in that, The step of inputting the concatenated feature vector into a preset decision model for forward computation, and outputting the personalized strategy vector through the decision model, includes: The decision model is a fully connected neural network, and the number of neurons in the input layer of the decision model is consistent with the dimension of the concatenated feature vector. The concatenated feature vector is input into the input layer of the decision model, and a nonlinear transformation is performed through at least one hidden layer of the decision model. A multidimensional policy vector is output at the output layer of the decision model. The first dimension of the strategy vector is used to characterize the dominant information output mode, the second dimension of the strategy vector is used to characterize the level of detail of the information broadcast, the third dimension of the strategy vector is used to index the predefined broadcast order template, and the fourth dimension of the strategy vector is used to control the activation status of the interaction channel.

4. The method according to claim 3, characterized in that, The step of dynamically selecting and serializing information entities in the structured information data according to the personalized strategy vector to generate an interactive information set includes: Based on the third dimension value in the personalized strategy vector, the corresponding broadcast order template identifier is searched in the meta-description layer of the structured information data; Based on the broadcast order template identifier, the semantic identifier sequence of the information entity to be processed is obtained from the order list defined in the meta-description layer; According to the order determined by the semantic identifier sequence, information entity data objects corresponding to each semantic identifier are extracted sequentially from the information entity layer of the structured information data. For each extracted information entity data object, based on the level of detail represented by the second dimension value in the personalized strategy vector, text content matching the level of detail is selected from the text detail mapping associated with the information entity data object as the core display content. Query whether the current information entity data object is registered as an interactive hotspot in the meta description layer. If so, generate an interactive context object based on the semantic identifier of the information entity data object and the predefined list of interactive actions in the meta description layer, and associate and bind the interactive context object with the core display content to form an information unit. All information units obtained through sequential processing are combined to generate the interactive information set.

5. The method according to claim 1, characterized in that, The generation of multimodal interaction instructions based on the interactive information set includes: Based on the arrangement order of each information entity in the interactive information set and its core display content, connect words based on natural language generation templates are inserted to generate a coherent voice broadcast text script. Based on the dimension values ​​in the personalized policy vector used to represent the dominant output modality and the real-time environment parameters in the multi-dimensional context parameters, the target speech rate is calculated using the following formula: in, For the target speech rate, As the system's baseline speech rate, This is the environmental factor adjustment coefficient. The ambient noise sound pressure level is obtained from the real-time environmental parameters. For reference noise sound pressure level, For user fitness adjustment coefficient, The sharpness compensation factor is calculated based on user historical interaction preference data; The multimodal interaction instructions that drive the local text-to-speech engine are generated. The multimodal interaction instructions include the speech-broadcast text script and the target speech rate. The multimodal interaction instructions are used to instruct the user device to synthesize and play audio.

6. A product information disclosure system based on QR codes, characterized in that, The system includes: The data acquisition module is used to acquire structured information data stored in the QR code graphic on the product packaging, wherein the structured information data includes key product information entities and meta-description data associated with the key product information entities; The personalized analysis module is used to analyze the structured information data based on multi-dimensional context parameters obtained locally from the user device, and generate a personalized strategy vector, wherein the multi-dimensional context parameters include at least one of user feature parameters, device capability parameters and real-time environment parameters obtained from the user device. The interaction strategy module is used to dynamically select and serialize the information entities in the structured information data according to the personalized strategy vector, and generate an interactive information set. The information disclosure module is used to generate multimodal interaction instructions based on the interactive information set, wherein the multimodal interaction instructions are used to instruct the user equipment to generate and output visual and / or auditory information from the structured information data.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 5.

Images