Commodity Recognition Using Image Feature Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional commodity recognition systems in canteen facilities are limited by their ability to recognize only specific containers, restricting their usability and increasing checkout time, as they require one-to-one correspondence with recognizable containers.
Innovation Solution
A commodity recognition apparatus using general object recognition technology, which captures images of food items, extracts feature data, and compares it to a recognition dictionary to identify items, allowing for flexible recognition without relying on specific container shapes or types, and integrates with IC-card systems for electronic payment and nutrition management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional recognition apparatus recognizes container on which commodities are placed, then recognition can be performed, but the container is limited to what is recognizable by the recognition apparatus, therefore the container is less likely to be used due to the limited freedom degree
Solution Approach 1:
The patent extracts the commodity recognition function from the container recognition function. Instead of recognizing the container as a whole, the system extracts and recognizes only the commodity items on the container through image processing, allowing any container type to be used without limiting the recognition capability
Solution Approach 2:
The recognition apparatus is designed to universally recognize any container type by focusing on the commodity items rather than the container itself. The system can handle various container shapes, sizes, and materials, making the container selection flexible while maintaining recognition accuracy
2Productivity
If cashier makes registration of the foods on the tray one by one in a cash register, then checkout can be performed, but it needs time for the checkout in the canteen facility
Solution Approach 1:
The system enables self-service checkout by automatically capturing images of the tray, extracting commodity information through image processing, and calculating the total amount without requiring cashier intervention. The customer simply places the tray on the recognition apparatus, and the system completes the checkout process autonomously
Solution Approach 2:
The manual mechanical process of cashier registration is replaced with an automated optical recognition system. The image acquisition module captures the tray contents, and image processing algorithms automatically identify and register commodities, eliminating the need for manual data entry and significantly speeding up checkout
Data Source
AI summary
A commodity recognition apparatus acquires an image including a commodity captured by an image capturing module and displays the acquired image on a display module. The commodity recognition apparatus displays a frame border surrounding the commodity on at least a portion of the image displayed on the display module. Moreover, the commodity recognition apparatus recognizes the commodity existing in the frame border according to a feature amount of the image in the area surrounded by the frame border. The commodity recognition apparatus outputs information of the commodity recognized.