Item Identification Using Brightness-Compensated Quantized Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current item identification processes at point of sale terminals are computationally intensive and memory-heavy, prone to misidentification due to variations in illumination, and vulnerable to fraud, as they rely solely on optical codes and compare large amounts of image data against extensive databases.
Innovation Solution
The process converts RGB image data to HSV format, uses weighted brightness values to compensate for illumination variations, and quantizes images to reduce data, employing a signal-to-noise ratio to efficiently match unknown items against a subset of known items, thereby reducing computational load and increasing identification accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If full-resolution image data is used for item identification, then identification accuracy is improved, but computational time and memory requirements increase significantly
Solution Approach 1:
The patent divides the image processing task into distinct stages: initial full-resolution capture for accuracy, followed by quantization to reduce resolution. This segmentation allows the system to benefit from both high-resolution data when needed and reduced data volume for efficient processing and database storage.
Solution Approach 2:
The patent applies quantization to change the resolution parameter of the image data, reducing it from full resolution to a lower resolution that retains sufficient identifying characteristics. This parameter change reduces the computational burden while maintaining adequate identification accuracy.
2Measurement precision
If full-resolution image data is stored in database, then item identification accuracy is improved, but memory requirements increase significantly
Solution Approach 1:
The patent segments the data storage requirement into two parts: storing quantized, lower-resolution images in the database for efficient retrieval, and using full-resolution data only when initially captured for comparison. This reduces the overall storage burden while maintaining identification capability.
Solution Approach 2:
The patent reduces the resolution parameter of stored image data through quantization, transforming full-resolution images into compact representations that occupy significantly less database space while preserving the essential visual characteristics needed for item identification.
3Device complexity
If unweighted brightness values are used in image comparison, then processing simplicity is maintained, but item misidentification rate increases due to illumination variations
Solution Approach 1:
The patent modifies the brightness parameter by applying weights during image comparison. This parameter change compensates for illumination variations by emphasizing or de-emphasizing certain brightness levels, thereby improving identification reliability without substantially increasing processing complexity.
4Productivity
If optical code scanning is used for item identification, then processing speed is improved, but fraud vulnerability increases
Solution Approach 1:
The patent merges optical code scanning with visual image recognition into a hybrid identification system. This combination maintains the speed advantage of code scanning while adding the fraud resistance of visual verification, as the system can detect discrepancies between the code and the actual item appearance.
Solution Approach 2:
The patent introduces visual image recognition as an intermediary verification layer between the optical code scan and the final identification. This intermediary step validates that the scanned code corresponds to the actual physical item, thereby preventing fraud while maintaining processing efficiency.
Data Source
AI summary
Methods and apparatus are provided for low resolution item identification using brightness compensated images. An image of an unknown item is captured in digital RGB image format and converted to HSV image format. Then the HSV formatted image is quantized which greatly reduces the resolution of the image. The quantized HSV image data of the unknown item is compared to a plurality of quantized HSV image data for known items. The comparison includes using a signal-to-noise ratio calculated using the quantized HSV image data for both the unknown and known items. A match is found when the calculated signal-to-noise ratio is above a predetermined threshold value.