Commodity information processing method and device, electronic equipment, storage medium and program product

Abstract

Embodiments of the present application provide a commodity information processing method and device, electronic equipment, storage medium and program product. The method responds to the acquisition of the after-sales request of the commodity, determines the target processing standard according to the type of the commodity corresponding to the after-sales request and the intention of the after-sales request, to determine the information acquisition guide required by the after-sales of the commodity, and the information to be processed acquired according to the target processing standard improves the qualified rate of the information to a certain extent; the target processing standard is sent to the conversation client corresponding to the after-sales request; in response to the acquisition of the information to be processed from the conversation client, a preset processing model corresponding to the information to be processed is called to verify the information to be processed, and in the case that the information to be processed is verified, the after-sales processing flow is executed, the automatic verification of the voucher is realized, the verification efficiency is improved, the long waiting time of the consumer is avoided, and the after-sales experience is improved.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a product information processing method, apparatus, electronic device, storage medium, and program product. Background Technology

[0002] In e-commerce service scenarios, when handling after-sales issues, service platforms typically require consumers to upload relevant credentials, such as product information, to support their after-sales requests. Verification of these credentials usually involves manual verification of their authenticity and whether they support the specific after-sales claim.

[0003] However, using manual verification of credentials can lead to a backlog of credentials during peak after-sales periods or when manual verification is offline, resulting in reduced verification efficiency, longer waiting times for consumers, and a poor after-sales experience. Summary of the Invention

[0004] To overcome the problems existing in the related technologies, this application provides a product information processing method, apparatus, electronic device, storage medium, and program product.

[0005] According to a first aspect of the embodiments of this application, a product information processing method is provided, the method comprising: In response to receiving an after-sales request, a target processing standard is determined based on the type of product corresponding to the after-sales request and the intent of the after-sales request; Send the target processing standard to the dialogue client corresponding to the after-sales request; In response to receiving pending information from the dialogue client, a preset processing model corresponding to the pending information is invoked to verify the pending information; the pending information is information received by the dialogue client from the user based on the target processing standard.

[0006] According to a second aspect of the embodiments of this application, a commodity information processing apparatus is provided, the apparatus comprising: The standard determination module is used to determine the target processing standard in response to receiving an after-sales request, based on the type of the product corresponding to the after-sales request and the intent of the after-sales request. The standard sending module is used to send the target processing standard to the dialogue client corresponding to the after-sales request; The verification module is used to respond to the information to be processed received from the dialogue client, and to call the preset processing model corresponding to the information to be processed to verify the information to be processed; the information to be processed is the information received by the dialogue client from the user based on the target processing standard.

[0007] According to a third aspect of the present application, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory, the computer program being executed by the processor to cause the electronic device to perform the method as described in the first aspect.

[0008] According to a fourth aspect of the embodiments of this application, a computer-readable storage medium is provided having a computer program stored thereon, the program being executed by a processor to implement the method as described in the first aspect.

[0009] According to a fifth aspect of the present application, a computer program product is provided, including instructions that, when executed on a computer, cause the computer to perform the method as described in the first aspect.

[0010] The technical solutions provided in this application embodiment may include the following beneficial effects: The product information processing method in this application embodiment, in response to receiving an after-sales request for a product, determines a target processing standard based on the type of product corresponding to the after-sales request and the intent of the after-sales request, thereby determining the information acquisition guidance required for the after-sales service of the product. The information to be processed obtained based on the target processing standard improves the information qualification rate to a certain extent. The target processing standard is sent to the dialogue client corresponding to the after-sales request. In response to receiving the information to be processed from the dialogue client, a preset processing model corresponding to the information to be processed is invoked to verify the information to be processed. This realizes automated verification of credentials, improves verification efficiency, avoids long waiting times for consumers, and enhances the after-sales experience.

[0011] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly described below. It should be understood that those skilled in the art can obtain other drawings based on these drawings without any creative effort.

[0013] Figure 1 This is a schematic diagram of a scenario corresponding to the product information processing method provided in the embodiments of this application; Figure 2 A schematic diagram of the server in the scenario corresponding to the product information processing method provided in the embodiments of this application. Figure 3 This is an exemplary flowchart corresponding to the product information processing method provided in the embodiments of this application; Figure 4This is an exemplary schematic diagram of the commodity information processing device provided in the embodiments of this application; Figure 5 An exemplary schematic diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0014] The technical solutions of the embodiments of this application will now be described with reference to the accompanying drawings.

[0015] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments and is not intended to limit the technical solutions of this application. As used in the specification and appended claims of this application, the singular expressions “a,” “an,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise.

[0016] It should also be understood that although the terms "first," "second," etc., may be used in the following embodiments to describe a certain type of object, the object should not be limited to these terms. These terms are used to distinguish the specific implementation objects of that type of object.

[0017] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in the embodiments of this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0018] The technical scenarios of the embodiments of this application are described below.

[0019] E-commerce platforms allow consumers to handle after-sales matters through online dialogue. For after-sales requests, consumers typically need to upload relevant supporting documentation. Verification of this documentation usually involves manual verification of its authenticity and whether it supports the after-sales requirement, so that the appropriate after-sales process can proceed if verification is successful.

[0020] However, the aforementioned method of manually verifying credentials relies on manual review of consumer-uploaded credentials, resulting in low efficiency. Furthermore, when human reviewers are offline or there is a backlog of credentials awaiting review, they are placed in a waiting queue until a human reviewer becomes available. This leads to reduced verification efficiency, longer waiting times for consumers, and a poor after-sales experience.

[0021] In view of this, embodiments of this application propose a product information processing method, which can be applied to a product information processing system. For example... Figure 1 As shown, this is a product information processing system that implements a product information processing method. Figure 1 As shown, the product information processing system includes a dialogue client 101 and a server 102, which are connected via a network. The dialogue client 101 can be hardware, such as a mobile phone, personal computer, tablet computer, wearable device, or other electronic device, or it can be an application (APP) or software module deployed on an electronic device. The dialogue client 101 serves as an online service portal for users. Users can log in to their online service portal through electronic devices or applications and send questions to the server 102 through this portal. The server 102 can be an application, service, instance, software module, virtual machine, container, or cloud server, or a hardware device or chip with data processing capabilities. After receiving an after-sales request, server 102 determines the target processing standard based on the type of product corresponding to the after-sales request and the intent of the after-sales request; sends the target processing standard to the dialogue client corresponding to the after-sales request; in response to obtaining the pending information from the dialogue client, it calls the preset processing model corresponding to the pending information to verify the pending information; the pending information is the information received by the dialogue client from the user based on the target processing standard; if the pending information is verified, the after-sales processing flow is executed.

[0022] In some embodiments, such as Figure 2 As shown, server 102 may include an application layer that depends on the upstream, an engine layer that depends on the algorithm, and an output layer. The application layer and the execution layer, as well as the execution layer and the engine layer, can interact with each other through an application programming interface (API).

[0023] The application layer is configured as the interaction medium between the after-sales dialogue platform and the user. The application layer can run on a terminal device. For example, it can deploy a dialogue client and a verification standard determination program for the information acquisition process. The dialogue client can receive user-input dialogue information in the form of a dialog box on the user interface (UI). The verification standard determination program is used to determine the information to be processed in the dialogue client's dialogue, so that after obtaining the information to be processed, relevant algorithms, models, and other modules in the engine layer can be called to verify it. In some embodiments, after obtaining an after-sales request, it is necessary to obtain the corresponding information to be processed. Different types of goods and different types of after-sales requests correspond to different information to be processed. To avoid inaccurate information uploaded by the user, the verification standard determination program can generate corresponding target processing standards. For example, for after-sales requests regarding unusable electronic goods and damaged fresh produce, unusable electronic goods typically require uploading video-type information of the product's operation or installation steps; while damaged fresh produce typically requires uploading image-type information of the product and its packaging.

[0024] The execution layer is configured as the coordination and processing decision-making layer of the after-sales dialogue interaction platform. The execution layer can carry the verification logic for information to be processed on the after-sales dialogue interaction platform. This verification logic can be a verification process, a verification agent, a verification program, etc. This application embodiment does not limit the type of the above verification logic; those skilled in the art can set it according to actual conditions.

[0025] The execution layer can maintain the calling interfaces, calling strategies, methods for determining execution components, methods for determining solutions, and a knowledge base for solutions related to the execution of the product information processing method. Execution components can be those corresponding to different stages such as "re-acquiring information to be processed," "transferring to manual review," and "verification count check," with each stage encapsulated as one or more execution components. The calling interface of the execution component defines the interaction protocol between components, such as the input / output data format, enabling the execution layer to call each execution component in a unified and pluggable manner. The calling strategy guides the execution layer on how to call execution components in specific scenarios. The strategy determines whether the execution order is serial, parallel, or conditional. For example, under what circumstances is it necessary to re-acquire information to be processed, or under what circumstances is it necessary to transfer to manual review? It also handles the scheduling of calls for execution components with dependencies, such as when one component needs to wait for the output of another component before it can run. The method for determining execution components is used to decide at runtime "which execution components to call" and "with what parameters to configure the call." This determination method is typically based on the identified target scenario, selecting a subset of target execution components from a pre-defined set of execution components. The solution knowledge base is used to store, manage, and retrieve solutions corresponding to historical scenarios.

[0026] The downstream engine layer is a capability center providing basic, general-purpose computing power and data services. It may also include image recognition modules, video analysis modules, natural language understanding modules, decision engine modules, and category-specific adaptive module libraries. The natural language understanding engine is used to understand user intent in historical dialogues to determine the intent of after-sales requests, and can also understand the text in the information to be processed to determine whether it matches the intent of the after-sales request. The image recognition and video analysis modules are used to identify and analyze the information to be processed. The image recognition module is used to perform image recognition on image-type information; the video analysis module is used to perform video analysis, specifically to extract keyframes representing the state of a product and perform action analysis, temporal analysis, etc., on the keyframes. The decision engine module is used to perform decision analysis based on the results of the image recognition, video analysis, and natural language understanding modules to determine whether the information to be processed has passed verification.

[0027] It should be understood that the above description of each functional layer of the commodity information processing system is merely illustrative and does not constitute a limitation on the functional layers of the commodity information processing system in this application embodiment. In actual implementation scenarios, the commodity information processing system may include more or fewer functional layers, and each layer may contain more or fewer functional modules. This application embodiment does not limit this.

[0028] It should be understood that Figure 1 and Figure 2The illustrated application scenario diagram is merely an illustrative representation of the commodity information processing system involved in the embodiments of this application and does not constitute a limitation on the technical solutions of the embodiments of this application.

[0029] In the above operating environment, the embodiments of this application provide as follows: Figure 3 The product information processing method shown is applied to, for example, [the following]. Figure 1 or Figure 2 The corresponding server is 102. The product information processing method includes the following steps: In step S301, in response to receiving an after-sales request, a target processing standard is determined based on the type of product corresponding to the after-sales request and the intent of the after-sales request.

[0030] For example, an after-sales request refers to a user's request submitted to an e-commerce platform or service system via a terminal device, requesting processing of purchased goods (such as refunds, exchanges, repairs, etc.). The product type refers to the categorization of sold goods according to certain classification standards. These standards can be an industry category system used internally by the e-commerce platform, such as primary categories (e.g., "fresh produce," "3C digital products," "clothing and footwear"), secondary categories (e.g., "fruits," "mobile phones," "outerwear"), or more granular classifications. Different types of goods have different physical characteristics, common quality issues, return and exchange policies, and key points for information verification.

[0031] The intent of an after-sales request refers to the core demand or nature of the problem claimed by the user through the request. Intent can be determined in various ways, such as through Natural Language Processing (NLP) analysis of the user's input text, or by obtaining information from a pre-defined list of options. Common after-sales intents include, but are not limited to, product spoilage / deterioration, scratches / damage, product malfunction / unusable, product not matching description, missing / absent items, etc.

[0032] The target processing standard refers to the information standard used to verify whether the information in an after-sales request complies with the information review rules. The information standard can be the information type, the amount of information, etc. The information type can be video or image, etc. This application embodiment does not limit the above target processing standard; those skilled in the art can set it according to actual circumstances.

[0033] Upon receiving a user's after-sales request, the same set of verification rules is not applied to all products and all types of after-sales issues. Instead, the system first determines the type of product involved in the request and the intent behind it. For example, does the user claim the product is rotten, scratched, malfunctioning, or physically damaged? The corresponding intent could be a refund, a partial refund, or a return. After determining the product type and the after-sales request, the system determines the target processing standard for that request from the information verification standards corresponding to that product type. The verification standard can be a decision tree model or a task rule engine. Using the product type and the after-sales intent as input conditions, the decision tree judges based on pre-defined branching logic to determine the target processing standard. For example, the first layer of the decision tree can branch based on the product type (e.g., "Is it fresh produce?"), and the second layer can branch based on the after-sales intent (e.g., "Is it a rotten issue?").

[0034] In step S302, the target processing standard is sent to the dialogue client corresponding to the after-sales request.

[0035] In step S303, in response to receiving the information to be processed from the dialogue client, a preset processing model corresponding to the information to be processed is invoked to verify the information to be processed.

[0036] Among them, the information to be processed is the information received by the dialogue client from the user based on the target processing standard.

[0037] For example, after determining the target processing standard, the target processing standard can be sent to the dialogue client, so that the user can receive the target processing standard through the dialogue client and then send the information to be processed that conforms to the target processing standard. The server then verifies the information to be processed after receiving it.

[0038] The pre-trained processing model refers to a machine learning model or algorithm module that is pre-trained and used to perform a specific verification task. Different types of goods can correspond to different verification models, and different types of information to be processed can also correspond to different types of verification models. For example, it may include, but is not limited to, a Convolutional Neural Network (CNN) model for detecting the spoilage of fresh produce, a target detection model for identifying scratches on 3C products, a temporal model for extracting keyframes from videos and analyzing operational norms, and a Generative Adversarial Network (GAN) detection model for identifying the authenticity of images, etc.

[0039] After receiving the information to be processed, based on the attributes and other information of the information to be processed, a preset processing model that matches the information to be processed is selected and invoked from a set of multiple pre-deployed verification models.

[0040] For example, if the information to be processed is an image of fresh produce, an image recognition model specifically trained for detecting spoilage in fresh produce will be invoked; if the information to be processed is a demonstration video of a 3C product, a video analysis model including keyframe extraction and action recognition will be invoked. The invoked verification model is responsible for conducting in-depth analysis of the information to be processed and outputting verification results, such as whether a product problem was detected, the type of problem, and its severity.

[0041] By dynamically calling the information to be processed, the matching and efficient utilization of the verification model are realized, which improves the accuracy and efficiency of verification to a certain extent.

[0042] As can be seen, by adopting the embodiments of this application, in response to obtaining an after-sales request for a product, a target processing standard is determined based on the type of product corresponding to the after-sales request and the intent of the after-sales request. This determines the information acquisition guidance for the standard required for after-sales service of the product. The information to be processed obtained based on the target processing standard improves the information qualification rate to a certain extent. The target processing standard is sent to the dialogue client corresponding to the after-sales request. In response to obtaining the information to be processed from the dialogue client, a preset processing model corresponding to the information to be processed is invoked to verify the information to be processed. This realizes automated verification of credentials, improves verification efficiency, avoids long waiting times for consumers, and enhances the after-sales experience.

[0043] In some embodiments, the target processing standard in step S301 above can be determined by: determining a set of candidate information standards that match the type of the product; filtering the target processing standard in the set of candidate information standards according to the intent of the after-sales request; the target processing standard includes information type; the information type includes at least one of images and videos.

[0044] For example, the candidate information standard set is a subset selected from all validation standard databases based on product type. This candidate information standard set contains all validation standard entries that may be related to the current product type. The candidate information standard set narrows the search scope and improves the matching efficiency of the target processing standard. For example, for fresh produce, the candidate information standard set may include spoilage detection standards, freshness standards, etc., and the corresponding information to be processed may include images of the product from multiple angles and images of the outer packaging, etc.; for clothing, the candidate information standard set may include damage detection standards, color difference standards, size mismatch standards, etc., and the corresponding information to be processed may be a global image and a partial detail image of the product, etc.

[0045] Information type is an attribute in the target processing standard, specifying the media format of the information to be processed by the user. Information types can be images, videos, or a combination of images and videos. Different information types correspond to different operation guidance methods and backend model invocation strategies. For example, when the information type is an image, the user is guided to take a photo and upload it, and an image recognition model is invoked for analysis; when the information type is a video, the user is guided to record and upload the video, and a video analysis model is invoked for processing.

[0046] In addition to specifying the information type of the information to be processed, the target processing standard may also include other attributes, such as the model identifier for verifying the information to be processed, the confidence threshold in the verification process, and the specifications for the image format and size of the information to be processed. This application does not limit the specific form of the above-mentioned target processing standard; those skilled in the art can set it according to actual circumstances.

[0047] In the implementation process, the candidate information standard set can be determined based on the hierarchical relationship corresponding to the existing category system of the e-commerce platform, and then the target processing standard can be further determined. Alternatively, the candidate information standard set can be determined based on other attribute types of the product. For example, the candidate information standard set can be determined based on whether the product belongs to fragile, perishable, or liquid-containing products, and then the target processing standard can be determined accordingly. This method is more flexible, not limited by fixed product categories, and the same type of product has the same verification standard, which also improves the accuracy of the target verification information.

[0048] In some embodiments, to reduce the cumbersome process of obtaining pending information, the pending information can be determined from historical dialogues first. For example, during the implementation process, it can be detected whether the historical dialogues corresponding to the after-sales request include initial pending information. If it is determined that the historical dialogues include initial pending information, the initial pending information is used as pending information. If it is determined that the historical dialogues do not include initial pending information, the target processing standard is sent to the client to obtain the pending information fed back by the dialogue client based on the target processing standard.

[0049] For example, historical dialogues are conversation messages sent by users through a dialogue client during after-sales requests. These historical dialogues may include request text, images, videos, and product order information. In this embodiment, after determining the target processing standard, it can first detect whether the user has previously uploaded the corresponding information. By scanning the historical dialogue information corresponding to the same product order, if initial pending information matching the information type in the target processing standard already exists in the dialogue information, it can be used as pending information. If no initial pending information matching the information type in the target processing standard exists in the dialogue information, the target processing standard needs to be further sent to the client to obtain the corresponding pending information.

[0050] During implementation, the information types, shooting angles, and shooting content requirements specified in the target processing standards can be converted into natural language text or encapsulated into standardized interactive component data (cards, forms, templates, etc.) and sent to the dialogue client. After parsing the data, the dialogue client renders a rich media interactive interface containing sample images, shooting instructions, upload buttons, and other elements, guiding the user to complete the credential upload according to the standard requirements.

[0051] If the information has already been uploaded in a previous conversation, the above judgment allows the information to be directly treated as pending information, avoiding duplicate uploads by users and improving user experience.

[0052] If the user has not yet uploaded the information to be processed, the user can be guided to upload the information according to the target processing standard. This will ensure that the image to be verified uploaded by the user in accordance with the target processing standard meets the target processing standard, and to a certain extent avoid the problem of repeated uploads caused by unqualified information to be processed.

[0053] It should be noted that the scope of historical conversations can be flexibly configured. For example, it can be limited to the history within the current session, or it can be extended to all conversation records of the user in the last N days. Those skilled in the art can set the retrieval scope and depth of historical conversations according to actual storage capacity and needs, and the embodiments of the present invention do not limit this.

[0054] In some embodiments, determining that a historical dialogue includes initial pending information means that the historical dialogue includes information of the same type as the information in the target processing standard, and that this information is used to support and substantiate after-sales requests. To further ensure the accuracy of the pending information, it can be confirmed with the user whether the information corresponds to and supports after-sales requests. In the implementation process, if initial pending information is detected in a historical dialogue, an inquiry message is sent to the dialogue client; if a confirmation message is received for the inquiry message, it is determined that the historical dialogue includes initial pending information. When sending the confirmation message, the corresponding pending confirmation information in the historical dialogue information can be referenced to send an inquiry message to the user, avoiding the user being unclear about what the pending confirmation information corresponds to. If a confirmation message is obtained, it can be used as pending information for further information verification.

[0055] Verification of information to be processed can determine different verification methods based on different information types to improve the accuracy of verification. In some embodiments, the information type is video. A preset processing model corresponding to the information to be processed is invoked to verify the information to be processed, including: invoking the video verification model to extract key frames representing the product status from the video credentials in the information to be processed; verifying whether the information to be processed matches the intent of the after-sales request based on the key frames; the intent of the after-sales request is determined based on historical dialogues.

[0056] For example, for information to be processed that is video, a video verification model can be invoked to verify the information. This video verification model can be a machine learning model or algorithm module specifically designed for processing and analyzing video content. The video verification model has the ability to process temporal information and can understand the dynamic content in the video that changes over time. It can include keyframe extraction modules, action recognition modules, temporal analysis modules, etc.

[0057] The video verification model can extract keyframes from a video and use these keyframes to determine whether the video matches the intent of the after-sales request. Keyframes can be single or multiple frames selected from a continuous frame sequence of the video that effectively represent the product's state or problem. Keyframes remove redundant footage irrelevant to the verification objective. Selection criteria for keyframes may include, but are not limited to, the clarity of the product in the image, the salience of the problem features, and the degree of change between frames.

[0058] For example, for fresh produce, in a video showing rotting fruit, the keyframe might be the frame where the rotten area is most clearly visible; for electronic products, if a user uploads a video showing Bluetooth headphones failing to charge, the keyframe could be a sequence frame of the Bluetooth headphones charging operation. The video verification model would automatically extract the moment when the user plugs the charging cable into the headphones but the indicator light doesn't turn on, or a close-up of the headphones not responding in the charging case.

[0059] In implementation, the video verification model calculates pixel or feature differences between adjacent frames. When the inter-frame difference exceeds a preset threshold, it is considered that a significant content change has occurred at that moment, and the frame is marked as a candidate keyframe. For example, when a user switches from displaying a perfectly good product to displaying a defective part, the inter-frame difference will increase significantly, and the frame at that switching moment will be extracted.

[0060] In some embodiments, when the information type is an image, an image verification model is invoked to determine whether the image in the information to be processed meets a preset information quality standard; wherein, the preset information quality standard includes at least one of the following: whether the information to be processed is a real product image, whether the information to be processed is a forged image, and whether the information to be processed is a reused image; if the image meets the preset information quality standard, the semantic features of the image are used to verify whether the information to be processed matches the intent of the after-sales request.

[0061] For example, an image verification model refers to a machine learning model or algorithm module used to verify static images. In this embodiment, the image verification model needs to possess two core capabilities: first, the ability to detect the quality of the image itself (such as determining authenticity or duplication); and second, the ability to semantically understand the image content (such as identifying product categories or detecting defect types). The image verification model can be a composite model integrating multiple functions, or it can be a cluster of models composed of multiple sub-models specializing in different tasks. This application embodiment does not limit the specific form of the image verification model; those skilled in the art can determine it according to the actual situation.

[0062] Product photos refer to images captured in real-time by a camera (such as a mobile phone camera) by the user regarding the product in question, taken after receiving the product but before applying for after-sales service. Fake images refer to images generated through technical means that are not actually taken. Fake images can include, but are not limited to: fake images generated by artificial intelligence image generation models (such as diffusion models), images obtained by altering real images using image editing software, and images downloaded from the internet that were not taken by the user. Reused images refer to images that are repeatedly sent or have been used in other after-sales applications or transaction disputes. Reused images may come from the same user's historical after-sales records or from other users' historical after-sales records. Detecting reused images is to prevent users from using the same photo of a problematic product to submit multiple orders or make multiple claims.

[0063] A model specifically designed for image verification is invoked to perform quality checks on the images in the information to be processed, determining whether the images meet the preset information quality standards. Only when the image passes all or part of the quality checks in this stage will the semantic matching stage proceed. If the image fails the checks, the information to be processed can be directly determined to be unqualified, and the corresponding information to be processed can be obtained again.

[0064] For images that pass the quality access verification, further semantic understanding is performed on the image content to extract information such as the main category of the product, the product's state (e.g., intact, damaged, rotten), and the type and severity of the problem. Then, these semantic features extracted from the image are compared with the after-sales request intent determined from historical dialogues to verify whether the two match. For example, if the after-sales intent is "strawberries are rotten," the image is checked to see if it actually contains one or more strawberries and whether there are visual features indicating rot on the strawberry's surface. If the semantic features match the intent, the verification passes; otherwise, it fails. Semantic features refer to visually understandable concepts extracted from the image content, such as "this is a strawberry," "the strawberry has white mold spots," and "the mold spots cover about one-third of the fruit's surface." Semantic feature extraction typically relies on deep learning models, especially Convolutional Neural Networks (CNNs) and Visual Transformer models.

[0065] The image verification process described above first checks the quality and authenticity of the image itself, filtering out unqualified credentials at the source. This avoids meaningless semantic analysis of invalid credentials later, saving computing resources. It also conducts a comprehensive evaluation from multiple dimensions, such as whether the image is a real object, whether it is a forged image, or whether it is a duplicate image, effectively preventing different types of fraud.

[0066] In some embodiments, the information to be processed may include both video information and image information. When determining whether a match is made, it is necessary to make a comprehensive judgment by combining these two types of information. For example, the first verification score of the video verification model and the second verification score of the image verification model are weighted and summed to obtain a fused verification score. If the fused verification score is greater than a preset score threshold, the information to be processed is verified.

[0067] For example, the verification score is a quantitative score output by the corresponding video or image verification model after analyzing the information to be processed. This verification score represents the confidence level of the verification model's output result for the information to be processed, and can be between 0 and 1 or between 0 and 100. Generally, a higher score indicates higher confidence and represents higher accuracy of the corresponding verification model. The preset score threshold is used to determine whether the information to be processed has passed verification. The preset score threshold can be set according to different scenarios. For example, a higher score threshold can be set for scenarios requiring manual review (higher risk) or multiple verifications, while a lower score threshold can be set for general scenarios. This application embodiment does not limit the specific value of the preset score threshold; those skilled in the art can set it according to actual conditions.

[0068] Two independent verification result scores are obtained from the video verification process and the image verification process, respectively. Preset weight coefficients are assigned to the first verification score and the second verification score, and then a weighted sum is performed to obtain a comprehensive "fusion verification score". For example, if the video credential is more reliable, a weight of 0.6 can be assigned to the first verification score and a weight of 0.4 to the second verification score, and then a weighted sum is calculated. Different weights can be assigned to different types of verification models, or different weights can be set according to different product types and verification scenarios. This application does not limit the above-mentioned method of assigning verification score weights; those skilled in the art can set them according to actual conditions.

[0069] The system compares the fused verification score with a pre-set threshold. If the fused verification score is greater than the threshold, the system determines that the overall verification of the information to be processed has passed; if the fused verification score is less than or equal to the threshold, the verification has failed, and the user may need to supplement the information to be processed or the case may be transferred to manual review. The above process achieves the aggregation and decision-making of multimodal information verification results by fusing the verification scores of video and image dimensions and making a final judgment based on the comparison of the fused score with the threshold.

[0070] In some embodiments, if the information to be processed passes verification, an after-sales processing procedure is executed; if the information to be processed fails verification, it is checked whether the number of verifications has reached a preset threshold; if the number of verifications has not reached the preset threshold, the information to be processed is re-obtained according to the target processing standard; if the number of verifications has reached the preset threshold, the information to be processed is transferred to manual verification.

[0071] For example, "verification passed" means that after the information to be processed is analyzed by a preset verification model, it shows that the product problem reflected in the information to be processed meets the conditions corresponding to the preset after-sales needs, and the product problem reflected in the information to be processed matches the intent of the after-sales request, and does not belong to false after-sales or malicious refund behavior.

[0072] The after-sales processing flow can be a standard operating procedure (SOP) for processing after-sales issues, provided that the information to be processed has been verified. In some embodiments, this after-sales processing flow may include generating a refund document and pushing it to the financial system to execute the refund, generating coupons and issuing them to the user's account, updating the order's after-sales status to "completed" or "refunded," sending an after-sales processing notification to the merchant, sending a processing result notification to the user (including information such as the refund amount and arrival time), and recording after-sales processing logs for auditing. The specific content and execution order of the after-sales processing flow can be configured differently based on factors such as product type, problem severity, and user level.

[0073] In some embodiments, during the execution of the after-sales processing flow, the processing nodes can be fed back to the user in real time. Specifically, a work order can be created, and the status of the work order represents the processing node. When the status of the work order changes, the corresponding node status is pushed to the user, so that the user can understand the after-sales processing flow in real time.

[0074] When the verification of the information to be processed fails, you can first check whether the current number of verification attempts exceeds the preset number threshold. The preset number threshold can be set to 3 times, which means that the user can try to upload credentials and accept verification a maximum of 3 times.

[0075] If the current verification count has not reached the preset threshold, it means that more information can be uploaded for processing. To ensure the accuracy of subsequent user-uploaded information, the target processing standard can be sent to the user again, guiding them to retake and re-upload the information according to that standard. Verification will then be performed again upon receiving the re-uploaded information. If the current verification count has reached the preset threshold, to improve after-sales efficiency, the user will no longer be guided to re-upload. Instead, the current information and related after-sales request context will be manually reviewed and processed.

[0076] As can be seen, by executing the above-mentioned product information processing method, in response to the acquisition of after-sales requests for products, the target processing standard is determined based on the type of product corresponding to the after-sales request and the intent of the after-sales request. This determines the information acquisition guidance for the standard required for after-sales service of the product. The information to be processed obtained based on the target processing standard improves the information qualification rate to a certain extent. The target processing standard is sent to the dialogue client corresponding to the after-sales request. In response to the acquisition of information to be processed from the dialogue client, the preset processing model corresponding to the information to be processed is invoked to verify the information to be processed. If the information to be processed passes the verification, the after-sales processing flow is executed. This realizes the automated verification of credentials, improves verification efficiency, avoids long waiting times for consumers, and enhances the after-sales experience.

[0077] The above implementation steps can also be implemented through software modules to achieve the corresponding functions. Corresponding to the above commodity information processing method, this application embodiment can also provide a commodity information processing device.

[0078] like Figure 4 As shown, a commodity information processing device is provided, which is applied to a target device. The device may include: a standard determination module 41, a standard sending module 42, a verification module 43, and an after-sales module 44.

[0079] For example: the standard determination module 41 is used to determine the target processing standard in response to receiving an after-sales request, based on the type of the product corresponding to the after-sales request and the intent of the after-sales request; The standard sending module 42 is used to send the target processing standard to the dialogue client corresponding to the after-sales request; The verification module 43 is used to respond to the information to be processed received from the dialogue client by calling a preset processing model corresponding to the information to be processed in order to verify the information to be processed; the information to be processed is the information received by the dialogue client from the user based on the target processing standard.

[0080] In some optional embodiments, the determination criteria module is further configured to: determine a set of candidate information criteria that match the type of the product; Based on the intent of the after-sales request, the target processing standard is filtered from the candidate information standard set; the target processing standard includes information type; the information type includes at least one of image and video.

[0081] In some optional embodiments, the verification module is further configured to: detect whether the historical dialogue corresponding to the after-sales request includes initial pending information; If it is determined that the historical dialogue includes the initial pending information, then the initial pending information is used as the pending information. If it is determined that the initial pending information is not included in the historical dialogue, a target processing standard is sent to the client to obtain the pending information fed back by the dialogue client based on the target processing standard.

[0082] In some optional embodiments, the verification module is further configured to: if the initial pending information is detected in the historical dialogue, send an inquiry message to the dialogue client; Upon receiving confirmation of the inquiry, it is determined that the historical dialogue includes the initial pending information.

[0083] In some optional embodiments, the information type is video, and the verification module is further configured to: invoke a video verification model to extract key frames representing the product status from the video credentials in the information to be processed; The keyframe is used to verify whether the information to be processed matches the intent of the after-sales request; the intent of the after-sales request is determined based on historical dialogues.

[0084] In some optional embodiments, the information type is an image, and the verification module is further configured to: call an image verification model to determine whether the image in the information to be processed meets a preset information quality standard; wherein the preset information quality standard includes at least one of the following: whether the information to be processed is a real product image, whether the information to be processed is a forged image, and whether the information to be processed is a reused image; If the image meets the preset information quality standard, the semantic features of the image are used to verify whether the information to be processed matches the intent of the after-sales request.

[0085] In some optional embodiments, when the information to be processed includes video information and image information, the verification module is further configured to: The first verification score of the video verification model and the second verification score of the image verification model are weighted and summed to obtain the fusion verification score. If the fusion verification score is greater than the preset score threshold, the information to be processed passes the verification.

[0086] In some optional embodiments, the device is further configured to: execute an after-sales processing procedure if the information to be verified passes verification; If the verification of the information to be processed fails, check whether the number of verification attempts has reached a preset threshold. If the number of verifications does not reach the preset threshold, the information to be processed is re-obtained according to the target processing standard. If the number of verifications reaches the preset threshold, the information to be processed will be transferred to manual verification.

[0087] As can be seen, by adopting the embodiments of this application, in response to obtaining an after-sales request for a product, a target processing standard is determined based on the type of product corresponding to the after-sales request and the intent of the after-sales request, thereby determining the information acquisition guidance for the standard required for after-sales service of the product. The information to be processed obtained based on the target processing standard improves the information qualification rate to a certain extent. The target processing standard is sent to the dialogue client corresponding to the after-sales request. In response to obtaining the information to be processed from the dialogue client, a preset processing model corresponding to the information to be processed is invoked to verify the information to be processed, thereby realizing automated verification of credentials, improving verification efficiency, avoiding long waiting times for consumers, and enhancing the after-sales experience.

[0088] Understandable Figure 4 The division of the various modules is merely a logical functional division. In actual implementation, the functions of these modules can be integrated into the hardware entity of the electronic device.

[0089] Please refer to Figure 5 , Figure 5 An electronic device is provided, and this disclosure also provides an electronic device for performing some or all of the steps of the above-described product information processing method. Please refer to... Figure 5 The electronic device includes a processor 500, a memory 501, a bus 502, and a communication interface 503, wherein the processor 500, the communication interface 503, and the memory 501 are connected via the bus 502; the memory 501 stores a computer program that can run on the processor 500, and the processor 500 can be equivalent to a CPU, executing the aforementioned provisions of this disclosure when running the computer program. Figure 3 The product information processing method provided in any of the embodiments, during the execution of the aforementioned method, interacts with the corresponding client through the communication interface 503 to exchange data and instructions.

[0090] The memory 501 may include high-speed random access memory (RAM) or non-volatile memory (NVM), such as at least one disk storage device. Communication between this system network element and at least one other network element is achieved through at least one communication interface 503 (which can be wired or wireless), such as the Internet, wide area network, local area network, metropolitan area network, etc.

[0091] Bus 502 can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. Memory 501 is used to store programs, and the processor 500 executes the programs after receiving execution instructions. Figure 2 The illustrated embodiments reveal that the commodity information processing method can be applied to or implemented by the processor 500.

[0092] The processor 500 may be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method can be completed by the integrated logic circuitry in the hardware of the processor 500 or by instructions in software form. The processor 500 may be a general-purpose processor, including a CPU, a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this disclosure. The general-purpose processor may be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this disclosure can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules may reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The storage medium is located in memory 501. The processor 500 reads the information in memory 501 and, in conjunction with its hardware, completes the steps of the above method.

[0093] The electronic device provided in this disclosure and the product information processing method provided in this disclosure are based on the same inventive concept and have the same beneficial effects as the methods they adopt, operate or implement.

[0094] This application also provides a computer-readable storage medium storing instructions for processing commodity information, which, when run on a computer, cause the computer to perform some or all of the steps in the method described in the foregoing embodiments.

[0095] This application also provides a computer program product that includes instructions for processing commodity information, which, when run on a computer, causes the computer to perform some or all of the steps in the method described in the foregoing embodiments.

[0096] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0097] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between apparatuses or units through some interfaces, and may be electrical, mechanical, or other forms.

[0098] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0099] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0100] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, smartphone, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0101] Although alternative embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make further changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0102] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of this application. It should be understood that the above description is only a specific embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, improvements, etc., made on the basis of the technical solution of this application should be included within the scope of protection of this invention.

Claims

1. A method for processing commodity information, characterized in that, The method includes: In response to receiving an after-sales request, a target processing standard is determined based on the type of product corresponding to the after-sales request and the intent of the after-sales request; Send the target processing standard to the dialogue client corresponding to the after-sales request; In response to receiving pending information from the dialogue client, a preset processing model corresponding to the pending information is invoked to verify the pending information; the pending information is information received by the dialogue client from the user based on the target processing standard.

2. The method according to claim 1, characterized in that, The step of determining the target processing standard based on the type of product corresponding to the after-sales request and the intent of the after-sales request includes: Determine a set of candidate information criteria that match the type of the product; Based on the intent of the after-sales request, the target processing standard is filtered from the candidate information standard set; the target processing standard includes information type; the information type includes at least one of image and video.

3. The method according to claim 1, characterized in that, The information to be processed obtained from the dialogue client includes: Detect whether the historical dialogues corresponding to the after-sales request include initial pending information; If it is determined that the historical dialogue includes the initial pending information, then the initial pending information is used as the pending information. If it is determined that the initial pending information is not included in the historical dialogue, a target processing standard is sent to the client to obtain the pending information fed back by the dialogue client based on the target processing standard.

4. The method according to claim 3, characterized in that, Determining that the historical dialogue includes the initial pending information includes: If the initial pending information is detected in the historical dialogue, an inquiry message is sent to the dialogue client; Upon receiving confirmation of the inquiry, it is determined that the historical dialogue includes the initial pending information.

5. The method according to claim 4, characterized in that, The information type is video. The step of calling a preset processing model corresponding to the information to be processed to verify the information includes: The video verification model is invoked to extract key frames representing the product status from the video credentials in the information to be processed. The keyframe is used to verify whether the information to be processed matches the intent of the after-sales request; the intent of the after-sales request is determined based on historical dialogues.

6. The method according to claim 4, characterized in that, The information type is an image. The step of calling a preset processing model corresponding to the information to be processed to verify the information includes: An image verification model is invoked to determine whether the image in the information to be processed meets a preset information quality standard; wherein, the preset information quality standard includes at least one of the following: whether the information to be processed is a real product image, whether the information to be processed is a forged image, and whether the information to be processed is a reused image; If the image meets the preset information quality standard, the semantic features of the image are used to verify whether the information to be processed matches the intent of the after-sales request.

7. The method according to claim 4, characterized in that, When the information to be processed includes video information and image information, verifying whether the information to be processed matches the intent of the after-sales request includes: The first verification score of the video verification model and the second verification score of the image verification model are weighted and summed to obtain the fusion verification score. If the fusion verification score is greater than the preset score threshold, the information to be processed passes the verification.

8. The method according to claim 1, characterized in that, The method further includes: If the information to be verified passes the verification, the after-sales processing procedure will be executed. If the verification of the information to be processed fails, check whether the number of verification attempts has reached a preset threshold. If the number of verifications does not reach the preset threshold, the information to be processed is re-obtained according to the target processing standard. If the number of verifications reaches the preset threshold, the information to be processed will be transferred to manual verification.

9. A commodity information processing device, characterized in that, The device includes: The standard determination module is used to determine the target processing standard in response to receiving an after-sales request, based on the type of the product corresponding to the after-sales request and the intent of the after-sales request. The standard sending module is used to send the target processing standard to the dialogue client corresponding to the after-sales request; The verification module is used to respond to the information to be processed received from the dialogue client, and to call the preset processing model corresponding to the information to be processed to verify the information to be processed; the information to be processed is the information received by the dialogue client from the user based on the target processing standard.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory, wherein the processor, when executing the computer program, implements the method of any one of claims 1-8.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by a processor to implement the method as described in any one of claims 1-8.

12. A computer program product, characterized in that, Includes instructions that, when executed on a computer, cause the computer to perform the method as described in any one of claims 1-8.

Images