Shelf Image Scale Correction via Deviation Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image recognition applications for shelf images face challenges in accurately detecting and identifying products due to inaccuracies in scale estimates of patterns and shelf images, leading to less accurate image recognition processes and the need for manual verification.
Innovation Solution
The implementation of an image processing system that determines optimized scales for patterns by using a median of multiple scale estimates and calculates a scale correction value based on deviation values between pattern sizes and object sizes in the shelf image, thereby improving the accuracy of scale estimates and image recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional image recognition applications are used to identify products in shelf images, then the system can detect and identify products, but the accuracy of product detection and identification is reduced due to inaccuracies in scale estimates
Solution Approach 1:
The system performs preliminary actions by calculating multiple scale estimates for each pattern before the actual image recognition process. These scale estimates are computed in advance using different reference objects and methods, allowing the system to select or combine the most accurate estimates for subsequent product detection, thereby improving both measurement precision and reliability
Solution Approach 2:
The system implements feedback by using detected objects in the shelf image to refine and recalculate scale estimates. The detected objects provide feedback information that is used to adjust the scale estimates of patterns, creating an iterative improvement process that enhances the accuracy of both scale estimation and product identification
2Measurement precision
If manual verification is performed to correct scale estimate errors, then the accuracy of image recognition can be improved, but the time and effort required for the audit process increases
Solution Approach 1:
The system performs self-service by automatically calculating multiple scale estimates and using them to correct its own scale estimation errors without requiring manual verification. The system uses detected objects to refine scale estimates and improve recognition accuracy autonomously, eliminating the need for time-consuming manual verification while maintaining high measurement precision
Solution Approach 2:
The system uses feedback from automatically detected objects to continuously refine scale estimates, creating a self-correcting mechanism that reduces the need for manual verification. This automated feedback loop maintains high accuracy while minimizing time loss by eliminating manual intervention
3Reliability
If multiple scale estimates are calculated for each pattern to improve accuracy, then the reliability of scale estimation is enhanced, but the device complexity and processing requirements increase
Solution Approach 1:
The system performs preliminary calculations of multiple scale estimates using readily available reference objects and standard measurement methods. By computing these estimates in advance before the main recognition process, the system enhances reliability without significantly increasing operational complexity, as the multiple estimates are generated using straightforward preliminary procedures
Solution Approach 2:
The system merges multiple scale estimates into a single optimized scale value by selecting or combining the most reliable estimates. This merging process simplifies the complex set of multiple estimates into a practical single value for use in image recognition, maintaining reliability while reducing the complexity of the estimation process
Data Source
AI summary
Example image processing methods, apparatus/systems and articles of manufacture are disclosed herein. An example apparatus includes an image recognition application to identify matches between stored patterns and objects detected in a shelf image, where the shelf image has a shelf image scale estimate. The example apparatus further includes a scale corrector to calculate deviation values between sizes of (A) a first set of the objects detected in the shelf image and (B) a first set of the stored patterns matched with the first set of the objects and reduce an error of the shelf image scale estimate by calculating a scale correction value for the shelf image scale estimate based on the deviation values.