Commodity identification method and device, storage medium and electronic device

By combining OCR text clustering and brand detection with product attribute matching, the problem of product recognition without image indexing was solved, achieving efficient product information matching and recommendation while reducing manual intervention.

CN115661516BActive Publication Date: 2026-06-09特赞(上海)信息科技有限公司
View PDF 3 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
特赞(上海)信息科技有限公司
Filing Date
2022-10-17
Publication Date
2026-06-09

AI Technical Summary

Technical Problem

When matching product image information in the system, existing technologies require pre-labeling of image information, which cannot effectively handle the massive number of cases without image indexing, resulting in a large workload for manual labeling.

Method used

By extracting the OCR text from the image to be uploaded and performing clustering, the product brand is detected. If no brand is detected, the brand name is matched, and further product attribute matching is performed, including weight, size and country of origin information. Finally, fuzzy matching is performed to identify the product.

Benefits of technology

It enables product recognition without indexed images, reduces manual labeling workload, improves recognition efficiency, and provides a recommended range of product categories to facilitate business personnel review.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115661516B_ABST
    Figure CN115661516B_ABST
Patent Text Reader

Abstract

The application discloses a commodity identification method and device, a storage medium and an electronic device, wherein the method comprises the following steps: in the case that a file name of a to-be-uploaded picture is not read and a target string of a commodity is not matched, extracting OCR text in the to-be-uploaded picture and clustering the OCR text to obtain long text with semantic information; detecting a brand of a commodity in the to-be-uploaded picture; if no brand is detected, matching a brand name according to the long text with semantic information; if there is a matched brand name, performing commodity attribute matching in a brand pool according to the long text with semantic information, wherein the commodity attribute matching at least comprises one of the following: weight information, size information and production country information; and performing fuzzy matching according to a commodity attribute matching result to obtain a final commodity identification result. Through the application, the commodity in a white background picture without a label can be well identified, and the accuracy of identification is ensured.
Need to check novelty before this filing date? Find Prior Art