Shopping Container Article Counting via Image Classification Weights
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting the number of articles in self-checkout systems, such as RFID tags and computer vision processing, are either costly and labor-intensive or unable to accurately determine the number of articles in a shopping container.
Innovation Solution
An apparatus and method using a detecting device to capture images of articles in a shopping container, a classifying device to classify the images using a classifier, and a calculating device to calculate a prediction value based on the detected number and classification weight, allowing for accurate detection of the number of articles at a relatively low cost.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If RFID tags are used to detect the number of articles, then the detection accuracy is improved, but the cost and labor intensity increase
Solution Approach 1:
The patent extracts only the necessary visual information from articles (presence, basic characteristics) rather than requiring full RFID tag data, achieving adequate detection accuracy without the high cost of universal RFID tagging
Solution Approach 2:
The system uses inexpensive image capture and processing instead of expensive RFID tags on each article, accepting that the detection method is simpler and less durable than RFID but sufficient for the application
2Ease of manufacture
If computer vision processing is used to determine shopping basket status, then the cost is reduced, but the measurement precision deteriorates
Solution Approach 1:
The patent segments the image processing into distinct stages: image acquisition, preprocessing, article detection, and counting. This segmentation allows each stage to be optimized independently, achieving better accuracy at lower cost
Solution Approach 2:
The system performs preliminary image preprocessing (enhancement, noise reduction) before article detection, improving the accuracy of subsequent detection steps while keeping overall costs low through efficient processing
3Ease of manufacture
If simple image detection is used to count articles, then the cost is reduced, but the reliability deteriorates
Solution Approach 1:
The patent implements feedback mechanisms where detection results are validated and cross-checked, and where system parameters are adjusted based on detection performance, improving reliability without increasing cost
Solution Approach 2:
The system dynamically adjusts detection parameters (thresholds, processing intensity) based on image quality and article characteristics, maintaining high reliability across varying conditions while keeping the system inexpensive
Data Source
AI summary
Embodiments of this disclosure provide an apparatus and method for detecting the number of articles and an electronic device. The apparatus for detecting the number of articles includes: a detecting device configured to detect the number of articles in a shopping container according to an image of articles in the shopping container, a classifying device configured to classify the image of the articles in the shopping container by using a classifier, and obtain a classification weight corresponding to a classification result; and a calculating device configured to calculate a prediction value of the number of the articles in the shopping container based on the number of the articles in the shopping container detected by the detecting device and the classification weight.