Dynamic Product Identification Dictionary Registration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing product identification systems for fruits and vegetables face decreased object recognition rates due to variations in texture and color across different regions of the same product type, leading to lower similarity scores when comparing captured images with pre-registered reference data.
Innovation Solution
The system includes a POS terminal with an imaging unit that captures images of products and performs dictionary registration, allowing for the extraction and storage of reference data associated with identification information, enabling new or additional registration of product features to improve recognition accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If reference data is registered in advance in a dictionary, then object recognition can be performed efficiently, but the object recognition rate decreases when products have different textures or colors in different regions
Solution Approach 1:
The patent makes the dictionary dynamic by allowing registration of multiple reference data items for the same product ID. Instead of a static single reference image, the system can now store multiple reference data items representing different textures and colors of the same product, enabling adaptive recognition that adjusts to variations in product appearance.
Solution Approach 2:
The patent changes the parameter of reference data from a single fixed value to multiple variable values. By allowing multiple reference data items with different texture and color parameters for the same product ID, the system can accommodate natural variations in product appearance while maintaining efficient recognition.
2Measurement precision
If multiple reference data items are registered for the same product, then recognition accuracy improves, but the complexity of the dictionary registration process increases
Solution Approach 1:
The patent enables automatic registration of multiple reference data items by capturing images of the same product from different positions or under different conditions. The system automatically processes and stores these images as separate reference data items without requiring manual intervention, thereby improving recognition accuracy while keeping the registration process simple.
Solution Approach 2:
The patent performs preliminary actions by automatically capturing and processing multiple images during the initial registration phase. By pre-capturing images from different positions and angles, the system prepares multiple reference data items in advance, eliminating the need for complex manual registration processes later.
3Adaptability or versatility
If the dictionary is updated dynamically with new reference data, then adaptability to product variations improves, but the system complexity increases
Solution Approach 1:
The patent implements a feedback mechanism where the system automatically captures images of products during normal operation and uses them to update the dictionary. The recognition results feed back into the system, allowing automatic registration of new reference data items that reflect actual product variations, thereby improving adaptability without requiring complex manual updates.
Data Source
AI summary
A product identification apparatus includes a storage device which stores a dictionary including reference data of each of a plurality of products stored in association with identification information of the product, an imaging section which captures image data of a target product, and a processor which performs a dictionary registration process. The processor extracts reference data of the target product from the image data captured by the imaging section. The storage device adds the extracted reference data to previously stored reference data in association with the identification information of the target product. Alternatively, the storage device stores the extracted reference data in association with the identification information of the target product when no reference data is previously stored in association with the identification information of the target product.