Product Recognition Apparatus Using Shelf-Based Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
As the number of products handled in a store increases, the processing load of a computer in recognizing products by collating between registered information and image-based information also increases, leading to inefficiencies in existing store systems.
Innovation Solution
A processing apparatus that acquires product pickup images, determines the product group on a shelf based on shelf-based display information, and recognizes products using recognition processing with a reduced set of feature values, thereby minimizing the processing load.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the number of products handled in a store increases, then the store's product variety and capacity increase, but the processing load of the computer in recognizing products increases
Solution Approach 1:
The patent segments the product recognition task by dividing products into groups based on their display locations (shelves). Instead of recognizing all products in the store, the system identifies which shelf a customer is at and only recognizes products on that specific shelf. This segmentation reduces the number of products to be processed at any given time while maintaining the ability to handle diverse product varieties across different shelves.
Solution Approach 2:
The patent applies local quality by tailoring the product recognition scope to the specific local context (shelf location) where the customer is present. The system adjusts the set of candidate products based on the customer's location, focusing recognition only on products relevant to that local shelf rather than processing all products in the entire store. This reduces processing load while maintaining accurate recognition for the local context.
2Adaptability or versatility
If the number of products as collation target increases, then the system can recognize more product types, but the processing load and time increase
Solution Approach 1:
The patent segments the collation target set by using shelf information to divide products into location-specific groups. The system determines which shelf the customer is at and only collates product information from that shelf, rather than considering all products in the store. This segmentation significantly reduces the number of collation targets and processing time while maintaining the ability to recognize diverse product types across different shelves.
Solution Approach 2:
The patent applies preliminary action by pre-determining which products are candidates for recognition based on shelf location before actual recognition processing. The system uses shelf-based display information to pre-filter and identify the relevant product group, so that when recognition needs to occur, only the pre-identified products need to be processed. This preliminary filtering action reduces processing time while maintaining comprehensive product recognition capability.
Data Source
AI summary
The present invention provides a processing apparatus (10) including: an acquisition unit (11) that acquires a product pickup image indicating a scene of picking up a product from a first product shelf by a customer; a determination unit (12) that determines a product group displayed on the first product shelf, based on shelf-based display information indicating a product displayed on each product shelf; and a first recognition unit (13) that recognizes a product included in the product pickup image by recognition processing in which the determined product group is set as a collation target among pieces of product feature value information indicating a feature value of an external appearance of a plurality of products.