Web Product Image Registration for Database Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for recognizing and registering products lack an efficient method for preregistering reference images, as none of the existing technologies provide a solution for this specific challenge.
Innovation Solution
A processing apparatus and method that acquires web product images based on search conditions, extracts product identification information, and registers these images and information in a database, enabling efficient registration of reference images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual registration of product reference images is used, then registration accuracy can be controlled, but registration efficiency is low and user workload is high
Solution Approach 1:
The system automatically acquires product images from web pages and extracts product identification information without requiring manual user input. The image acquisition unit fetches images from URLs obtained through web searches, and the extraction unit automatically identifies product information, eliminating the need for users to manually upload and label images.
Solution Approach 2:
The system performs preliminary web searches to collect product images and information before the actual registration process. By proactively acquiring images from multiple web pages and pre-extracting product identification information, the system prepares data in advance, reducing the overall time and effort required for registration.
2Adaptability or versatility
If multiple web pages are searched for product images, then image selection flexibility increases, but processing time increases
Solution Approach 1:
The system performs web searches with predetermined search conditions to acquire a set of candidate images from multiple web pages. By collecting more images than strictly necessary and then selecting appropriate ones, the system ensures sufficient image options while managing processing time through efficient extraction and verification processes.
Solution Approach 2:
The system conducts web searches and collects product images from multiple pages in advance, before final selection is needed. By pre-acquiring and pre-processing images from various sources, the system prepares a pool of candidate images that can be quickly selected and verified, reducing overall processing time.
3Productivity
If automatic extraction of product identification information is implemented, then registration speed increases, but information accuracy may decrease
Solution Approach 1:
The extraction unit automatically extracts product identification information from web pages, and the verification unit provides feedback by checking the extracted information against the acquired images and search results. This feedback mechanism ensures that the extracted information is accurate before registration, combining automatic processing with verification.
Solution Approach 2:
The system performs self-verification of extracted product identification information by cross-checking against the acquired images and search query results. The verification unit automatically confirms the accuracy of extracted data, eliminating the need for manual verification while maintaining high accuracy standards.
Data Source
AI summary
The present invention provides a processing apparatus (10) including: a web product image acquisition unit (11) acquiring a web product image publicly available in a web page determined based on a predetermined search condition; a product identification information acquisition unit (12) acquiring product identification information of a product indicated by the web product image; and a registration unit (13) registering the web product image and the product identification information in association with each other in a product database.