Commodity attribute AI generation method based on national standard bar code and big data

Abstract

The invention relates to the technical field of electronic commerce information processing, in particular to a commodity attribute AI generation method based on national standard bar codes and big data, and the method comprises the steps: collecting multi-source data, constructing a dynamically updated commodity knowledge graph, and forming a standardized commodity basic information base; then, taking a national standard bar code as a key index, realizing intelligent identification and automatic filling of commodity information by using an AI technology, and having a self-learning ability for unknown commodities; further, carrying out AI optimization on commodity attributes, including complementation, error correction and specification standardization; in addition, standardized commodity main picture and detail page content can be automatically generated, and AI auditing is carried out on the standardized commodity main picture and detail page content to ensure compliance and quality. According to the invention, the whole process automation of commodity information from collection, standardization to application is realized, the labor cost is greatly reduced, the data quality and the circulation efficiency are improved, and a unified and reliable data base is provided for constructing intelligent retail ecology.

Description

Technical Field

[0001] This invention relates to the field of e-commerce information processing technology, specifically to an AI-based method for generating product attributes based on national standard barcodes and big data. Background Technology

[0002] With the deepening development of e-commerce and the retail industry, the digital management of product information has become a core bottleneck for improving industry efficiency. Currently, the industry generally faces two major pain points: "data shortage" and "data chaos". Data shortage: A massive amount of product information needs to be manually entered, which is costly and inefficient, and there is a serious lack of information for new products and long-tail products.

[0003] Data inconsistency: The descriptions of the same product vary greatly across different platforms and merchants, and the specifications are inconsistent, such as "500ml", "500 milliliters", and "0.5L" coexisting. This leads to inaccurate search engines, difficulty in price comparison, impaired consumer experience, and high return rates.

[0004] While existing technologies offer solutions for product retrieval using barcodes, they are mostly limited to simple database matching, unable to handle products unknown to the system, and suffer from single data sources, outdated updates, and a lack of in-depth mining, standardized reconstruction, and intelligent generation capabilities for product information.

[0005] As digital transformation enters a more complex phase, there is an urgent need for an infrastructure-level solution that can automatically, in real-time, and in a standardized manner generate and manage product attributes.

[0006] Therefore, those skilled in the art have provided a product attribute AI generation method based on national standard barcodes and big data to solve the problems mentioned in the background art. Summary of the Invention

[0007] To address the aforementioned technical issues, this invention provides a product attribute AI generation method based on national standard barcodes and big data. This method uses national standard barcodes as a unified entry point and comprehensively utilizes AI technologies such as big data fusion, knowledge graphs, natural language processing (NLP), and computer vision (CV) to construct a self-learning and self-evolving digital foundation for products, thereby achieving full-process automation from product information collection, standardization, and intelligent filling to content generation and review.

[0008] A product attribute AI generation method based on national standard barcodes and big data includes the following steps: S1. Construct a dynamically updated basic product information database: Through multi-source data collection, big data fusion and cleaning, construct and continuously update a structured product knowledge graph; S2. AI-powered intelligent identification and product information creation for merchants: Receives product identifier input from merchants, performs intelligent matching and automatic filling based on the product basic information database, and processes unknown products to achieve system self-learning; S3. AI optimization of basic product attributes and specifications: Complete, correct and standardize product attributes, and recommend standardized attribute values ​​to merchants; S4. AI generation of product images and product detail pages: Automatically generate standardized main product images and product detail page text and templates based on product attributes; S5. AI review of product images and text content: Review the compliance, quality and consistency of the generated images and text content.

[0009] Preferably, the construction of the product knowledge graph in step S1 specifically includes: Product data was collected from the official database of national standard barcodes, publicly available data from e-commerce platforms, and industry standard documents; The collected data is cleaned, deduplicated, and correlated, and product information from different sources is aligned using the national standard barcode as the key index; Construct a graph structure with products, brands, categories, and attributes as entities, and define the relationships between them. The graph clearly defines the set of standard attributes corresponding to each product category.

[0010] Preferably, the AI-powered intelligent recognition and product information creation for merchants in step S2 specifically includes: It supports receiving product identification via at least one of the following methods: barcode scanning, product image upload, or text input; The identifier is matched in the product knowledge graph. If successful, the structured attributes are automatically filled into the merchant form. If a match fails, natural language processing and computer vision technologies are used to extract product attributes from the input information to form a confirmation form. After confirmation and review, the new product information is entered back into the product basic information database.

[0011] Preferably, step S3 includes: By using AI models to compare information from multiple sources, missing attribute values ​​for products can be automatically filled in and erroneous attribute values ​​can be corrected. Identify non-standard descriptions of product specifications and convert them into a preset standard format; When merchants are editing, standard attribute fields and commonly used values ​​are recommended from the knowledge graph based on the product category.

[0012] Preferably, step S4 includes: Perform subject recognition, background removal, and background replacement on product images, and automatically add marketing tags that meet the specifications; Based on the product's fundamental attributes, a large natural language processing model is used to automatically generate product titles, selling point descriptions, and product detail text, and automatically format and generate a product detail page template.

[0013] Preferably, step S5 includes: Identify and inspect product images for violations, and verify the consistency between the image content and the product category and title. We conduct filtering of prohibited words, standardization checks, and intellectual property risk screening of product text content.

[0014] The technical effects and advantages of this invention are as follows: For businesses and merchants: It enables "one-click creation" of product information, reducing manual data entry costs by more than 75%, greatly shortening the product listing cycle, and automatically adapting to multiple platform rules, reducing repetitive operations.

[0015] For the industrial chain: an industry-level, standardized commodity information database has been built, breaking down data silos and providing a reliable data foundation for supply chain collaboration and intelligent inventory management, which is expected to increase inventory turnover by about 20%.

[0016] For consumers: It provides accurate, consistent, and comprehensive product information, significantly improving search matching and shopping experience, and reducing return rates due to inconsistent information.

[0017] For platforms and the industry: AI review ensures the compliance and quality of content, and the accumulated data assets can be used to benefit the industry, driving the entire retail ecosystem towards greater efficiency, intelligence, and transparency. Attached Figure Description

[0018] Figure 1 This application provides an overall system architecture for a product attribute AI generation method based on national standard barcodes and big data, as illustrated in this embodiment. Figure 2 This is a detailed architecture diagram of a product attribute AI generation method based on national standard barcodes and big data provided in an embodiment of this application; Figure 3 This is a detailed flowchart of the AI ​​intelligent recognition and product creation process for merchants in a product attribute AI generation method based on national standard barcodes and big data provided in this application embodiment. Detailed Implementation

[0019] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments. The embodiments of the present invention are given for illustrative and descriptive purposes only, and are not intended to be exhaustive or to limit the invention to the forms disclosed. Many modifications and variations will be apparent to those skilled in the art. The embodiments were chosen and described to better illustrate the principles and practical application of the invention, and to enable those skilled in the art to understand the invention and design various embodiments with various modifications suitable for a particular purpose.

[0020] Example 1

[0021] Please see Figures 1-3 This embodiment provides a product attribute AI generation method based on national standard barcodes and big data, including... S1. Construct a dynamically updated basic product information database: Through multi-source data collection, big data fusion and cleaning, construct and continuously update a structured product knowledge graph; in, Multi-source data collection: The system will automatically capture and aggregate commodity data from multiple channels, including: the underlying database of national standard barcodes, such as GS1 China Article Numbering Center, to obtain legal identification information such as manufacturer codes and commodity classifications; Public data sources: publicly available descriptions from e-commerce platforms, user reviews, industry standard documents, category directories, etc. Big data fusion and cleaning: Utilizing big data technology to clean, deduplicatize, denoise, and correlate heterogeneous data from different channels to solve the problem of inconsistent descriptions of the same product on different platforms; Knowledge Graph Construction: The system does not simply store data, but builds a structured product knowledge graph. In this graph, entities such as "brand", "category", and "attribute" and their relationships such as "belong to" and "own" are clearly defined. For example, the graph will define the category "mobile phone" and have attributes such as "brand", "model", "screen size", and "memory".

[0022] Dynamic update mechanism: Establish a continuous learning mechanism. When new products are launched in the market or product information is updated, the system can automatically detect, capture and update the basic information database to ensure the timeliness of the data.

[0023] S2. AI-powered intelligent identification and product information creation for merchants: AI-powered intelligent identification and product information creation for merchants: Receives product identifier input from merchants, performs intelligent matching and automatic filling based on the product basic information database, and processes unknown products to achieve system self-learning; Among them, this is the core application function for merchants, which can create products with "one click"; Diverse input methods: Barcode scanning: Merchants use barcode scanners or mobile phones to scan product barcodes; Image recognition: Merchants upload images of product packaging, and the system automatically identifies and extracts the main product and text using CV technology; Text input: Merchants enter the product name or part of the keywords; Smart Match and Fill: The system uses barcodes or recognition results as "keys" to perform millisecond-level matching in a vast database of basic product information; Once a match is successful, the structured basic product attributes, such as brand, product name, specifications, and net content, will be automatically populated into the merchant's backend product form.

[0024] Handling "Unknown Products": When scanning a barcode unknown to the system, the system uses NLP and CV technologies to try to extract key attributes from the images and text uploaded by the merchant and prompts the merchant for confirmation. After confirmation, the new product information can be reviewed and then added back to the basic information database to achieve self-learning evolution.

[0025] S3. AI optimization of basic product attributes and specifications: Complete, correct and standardize product attributes, and recommend standardized attribute values ​​to merchants; This feature makes product information more accurate, richer, and more valuable; Attribute completion and error correction: Even data in the basic information database may be missing or incorrect; the system uses an AI model to automatically complete missing attributes by comparing information from multiple sources, such as adding "material" and "flavor", and correcting existing erroneous information. Specification standardization: Non-standard specification descriptions, such as "500ml", "500 milliliters", and "0.5L", are uniformly converted into a standard format, such as "500ml", to facilitate machine processing and consumer comparison; Attribute value recommendation: When merchants manually edit, the system will recommend the most commonly used attribute fields and values ​​in the category based on the product category, guiding merchants to input standardized information and avoiding arbitrary input.

[0026] S4. AI generation of product images and product detail pages: Automatically generate standardized main product images and product detail page text and templates based on product attributes; Among these features, the extension of this technology improves the efficiency of merchants' marketing efforts; Intelligent main image generation: Background optimization: Automatically identify the main product image, cut out the image, and replace it with a solid color or a standardized background that matches the brand's tone; Information synthesis: Based on key product attributes, such as "new product" or "hot selling", automatically add compliant marketing tags or text watermarks to images; AI-powered product detail page creation: Automatic copywriting generation: Based on the product's basic attributes, such as brand, name, ingredients, and functions, it uses NLP models to automatically generate attractive product titles, selling point descriptions, and detailed copy. Automatic layout: Based on the generated text and existing image materials, it automatically generates a clear and visually appealing details page template, greatly reducing the workload of graphic designers and copywriters.

[0027] S5. AI review of product images and text content: Review the compliance, quality and consistency of the generated images and text content; This is to ensure that the generated content is of high quality and compliant; Image review: Violation Content Detection: Automatically identifies whether images contain pornographic, violent, or politically sensitive content. Quality inspection: Check whether the image is blurry, distorted, or contains irrelevant watermarks or contact information; Content consistency review: Compare whether the product images match the title and category. For example, the image is of shampoo, but the category is selected as "food". Text content review: Prohibited / Sensitive Word Filtering: Automatically detects whether there are absolute terms prohibited by advertising law in the title and details page, such as "best quality", "number one brand", false advertising words or sensitive information; Compliance review: Check whether the brand, specifications and other information are filled in correctly and in accordance with the platform's publishing rules; Preliminary intellectual property screening: Through text comparison, it provides early warnings of potential brand trademark abuse and infringing descriptions.

[0028] Examples of the present invention: I. The system of the present invention includes a data acquisition layer, a data processing and knowledge graph layer, and an application service layer.

[0029] At the data acquisition layer, raw data is obtained from GS1 encoding centers, mainstream e-commerce platform APIs, and other channels through a crawler cluster.

[0030] At the data processing and knowledge graph layer, big data frameworks such as Spark are used for data cleaning and fusion, and graph databases such as Neo4j are used to build and maintain the product knowledge graph. For example, define inherent attributes for the "smartphone" category such as "brand", "model", "screen size", "RAM", etc., and establish a relationship such as "iPhone 14 has a 6.1-inch screen".

[0031] At the application service layer, the services described in S2 to S5 are provided to the outside world through a microservice architecture.

[0032] 2. When a merchant scans the barcode of a snack product with their mobile phone, the system matches the product in the knowledge graph and automatically fills in attributes such as "Brand: XX", "Product Name: Chocolate Flavored Wafer" and "Net Weight: 150g" into the backend form. The system finds that the product is missing the "Flavor" attribute, so it analyzes user reviews to find "Chocolate Flavor" and automatically completes it. At the same time, the system standardizes the specification "150g" to "150g". Finally, it generates a main image with a "New Product Launch" label and a details page copy highlighting the "Rich Chocolate" selling point through S4.

[0033] When a merchant scans an imported product that is unknown to the system, the system extracts key information such as brand, product name, and specifications by recognizing the text on the product packaging image, and generates a product form to be confirmed. After the merchant provides supplementary confirmation, the product information is reviewed and approved by the auditor, and then added to the knowledge graph as a new entry.

[0034] In summary, this invention, through innovative combination of technologies, upgrades the universal national standard barcode into an intelligent data entry point, constructing a dynamic, standardized, and intelligent digital foundation for commodities. It effectively solves the long-standing data governance problem in the industry, demonstrating significant technological advancement and broad industrial application value.

[0035] Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art and related fields based on the embodiments of the present invention without inventive effort should fall within the scope of protection of the present invention. Structures, devices, and operating methods not specifically described and explained in the present invention, unless otherwise specified or limited, shall be implemented according to conventional means in the art.

Claims

1. A product attribute AI generation method based on national standard barcodes and big data, characterized in that, Includes the following steps: S1. Construct a dynamically updated basic product information database: Through multi-source data collection, big data fusion and cleaning, construct and continuously update a structured product knowledge graph; S2. AI-powered intelligent identification and product information creation for merchants: Receives product identifier input from merchants, performs intelligent matching and automatic filling based on the product basic information database, and processes unknown products to achieve system self-learning; S3. AI optimization of basic product attributes and specifications: Complete, correct and standardize product attributes, and recommend standardized attribute values ​​to merchants; S4. AI generation of product images and product detail pages: Automatically generate standardized main product images and product detail page text and templates based on product attributes; S5. AI review of product images and text content: Review the compliance, quality and consistency of the generated images and text content.

2. The method for generating product attributes using AI based on national standard barcodes and big data according to claim 1, characterized in that, The construction of the product knowledge graph in step S1 specifically includes: Product data was collected from the official database of national standard barcodes, publicly available data from e-commerce platforms, and industry standard documents; The collected data is cleaned, deduplicated, and correlated, and product information from different sources is aligned using the national standard barcode as the key index; Construct a graph structure with products, brands, categories, and attributes as entities, and define the relationships between them. The graph clearly defines the set of standard attributes corresponding to each product category.

3. The method for generating product attributes using AI based on national standard barcodes and big data according to claim 1, characterized in that, The AI-powered intelligent recognition and product information creation for merchants in step S2 specifically includes: It supports receiving product identification via at least one of the following methods: barcode scanning, product image upload, or text input; The identifier is matched in the product knowledge graph. If successful, the structured attributes are automatically filled into the merchant form. If a match fails, natural language processing and computer vision technologies are used to extract product attributes from the input information to form a confirmation form. After confirmation and review, the new product information is entered back into the product basic information database.

4. The method for generating product attributes using AI based on national standard barcodes and big data according to claim 1, characterized in that, Step S3 includes: By using AI models to compare information from multiple sources, missing attribute values ​​for products can be automatically filled in and erroneous attribute values ​​can be corrected. Identify non-standard descriptions of product specifications and convert them into a preset standard format; When merchants are editing, standard attribute fields and commonly used values ​​are recommended from the knowledge graph based on the product category.

5. The method for generating product attributes using AI based on national standard barcodes and big data according to claim 1, characterized in that, Step S4 includes: Perform subject recognition, background removal, and background replacement on product images, and automatically add marketing tags that meet the specifications; Based on the product's fundamental attributes, a large natural language processing model is used to automatically generate product titles, selling point descriptions, and product detail text, and automatically format and generate a product detail page template.

6. The method for generating commodity attributes using AI based on national standard barcodes and big data according to claim 1, characterized in that, Step S5 includes: Identify and inspect product images for violations, and verify the consistency between the image content and the product category and title. We conduct filtering of prohibited words, standardization checks, and intellectual property risk screening of product text content.

Images