Product Feature Value Updating for Accurate POS Object Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In point-of-sale systems, the accuracy of object recognition for products without symbol codes, such as vegetables and fruits, is compromised when feature values need to be updated due to limited storage space, as random replacement of feature data can decrease recognition accuracy.
Innovation Solution
An information processing apparatus with a storage unit, image capturing unit, and processing unit that selects and replaces feature values based on classification and usage frequency, ensuring high similarity and maintaining recognition accuracy by using a replacement method appropriate for each product.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If new feature values are added to maintain recognition accuracy, then recognition accuracy is improved, but storage space is consumed
Solution Approach 1:
The system discards existing feature values from storage and recovers storage space by selectively replacing them with new feature values. The selection is based on criteria such as oldest update date, random selection, or lowest usage frequency, ensuring that discarded feature values are those least likely to impact recognition accuracy.
Solution Approach 2:
The system changes the parameter of feature value selection by introducing multiple selection criteria (oldest update date, random selection, lowest usage frequency) to determine which feature values to replace. This allows flexible adaptation to different storage constraints and recognition requirements.
2Quantity of substance
If feature values are replaced to maintain storage capacity, then storage space is preserved, but recognition accuracy deteriorates
Solution Approach 1:
The system recovers storage space by replacing feature values while minimizing impact on recognition accuracy. It discards feature values that are least valuable (determined by selection criteria) and replaces them with new feature values, thus recovering storage without significantly degrading recognition performance.
Solution Approach 2:
The system performs preliminary evaluation of feature values using selection criteria (oldest update date, random selection, or lowest usage frequency) before replacement. This preliminary action identifies which feature values can be safely replaced, preventing deterioration of recognition accuracy.
3Ease of operation
If random replacement of feature data is performed, then storage management is simplified, but recognition accuracy decreases
Solution Approach 1:
The system changes the parameter of feature value selection from purely random to multi-criteria-based selection. By introducing selection criteria (oldest update date, random selection, or lowest usage frequency), the system maintains ease of operation while significantly improving recognition accuracy by making informed replacement decisions.
Solution Approach 2:
The system performs self-service by automatically evaluating and selecting feature values for replacement based on predetermined criteria. This eliminates the need for complex manual intervention while maintaining high recognition accuracy through intelligent selection.
Data Source
AI summary
An information processing apparatus includes a storage unit, an image capturing unit, and a processing unit. The storage unit stores a plurality of feature values to be used for object recognition and an update program for the feature values, with respect to each of products registered for sale. The image capturing unit is configured to acquire an image of a product registered for sale. The processing unit is configured to extract a feature value of the product from the acquired image, select one of the plurality of the feature values corresponding to the product as a replacement target, by executing the update program corresponding to the product, and replace the selected feature value with the extracted feature value.