Shelf Image Recognition Using Adjacent-Article Inference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing systems struggle to accurately infer the names of unrecognized articles with high precision.
Innovation Solution
An image processing apparatus and method that includes an image processing unit to determine article names and an article inference unit to infer the names of undetermined articles based on specific conditions, such as adjacent article names or shelf positioning, when the initial determination fails.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If image processing is used to determine article names, then automation is improved, but measurement precision deteriorates for unrecognized articles
Solution Approach 1:
The patent introduces an inference unit as an intermediary component that bridges the gap between image processing results and final article name determination. When the image processing unit fails to recognize an article, the inference unit steps in to determine the article name by analyzing spatial relationships with adjacent recognized articles, thereby maintaining both automation and precision.
Solution Approach 2:
The system implements a feedback mechanism where the image processing unit's output is evaluated, and when recognition fails, the result is fed back to the inference unit for alternative processing. This feedback loop ensures that unrecognized articles receive additional processing attention, improving overall measurement precision while maintaining automation.
2Measurement precision
If inference based on adjacent articles is used, then measurement precision is improved for unrecognized articles, but device complexity increases
Solution Approach 1:
The patent segments the article name determination process into two distinct functional modules: an image processing unit for direct recognition and an inference unit for indirect determination. This segmentation allows each unit to specialize in its strength while keeping the overall system complexity manageable through clear functional separation.
Solution Approach 2:
The inference unit is activated only partially - specifically when the image processing unit fails to recognize an article. This partial action approach avoids the complexity of continuously running both systems, applying the more complex inference method only when necessary, thus improving precision for unrecognized articles without excessively increasing overall system complexity.
Data Source
AI summary
By processing an image in which a plurality of articles (such as products) placed on a shelf are captured, an image processing unit (120) determines the article name (such as the product name) of each of the plurality of articles captured in the image. Determining an article name also includes determining identification information tied to the article name. When a specific condition is satisfied for an article the article name of which is indeterminable by the image processing unit (120) (hereinafter described as an undetermined article), an article inference unit (130) infers the article name of the undetermined article to be the article name of an article positioned adjacent to the undetermined article.