Low-Power Self-Checkout Device Using Saliency-Based Product Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional smart shopping carts rely on code scanners for product identification, which are limited in functionality and efficiency, especially in terms of power consumption and processing capabilities.
Innovation Solution
A low-power self-checkout shopping device equipped with image sensors, computer hardware, and a multi-state recognition pipeline that includes a saliency detector, change detector, feature detector, feature recognizer, product database, and UI display, allowing for efficient product identification and registration without the need for continuous connectivity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If code scanners are used for product identification, then the device is simple and low-cost, but the functionality and processing capabilities are limited
Solution Approach 1:
The shopping cart integrates multiple functions including image sensing, barcode scanning, QR code recognition, product identification, and checkout processing into a single device. The image sensor serves multiple purposes: capturing product images for identification, detecting product placement, and providing visual feedback to users, thereby increasing versatility without proportionally increasing complexity
Solution Approach 2:
The patent combines the code scanner functionality with image sensor capabilities into a unified processing system. The computer hardware integrates both barcode/QR code scanning and image-based product recognition algorithms, merging previously separate identification methods into a single multi-functional module that processes various product identification tasks simultaneously
2Measurement precision
If continuous image processing is performed to ensure accurate product recognition, then recognition accuracy is improved, but power consumption increases
Solution Approach 1:
The system performs image processing periodically rather than continuously. The image sensor captures images at predetermined time intervals, and the computer hardware processes only these periodic images for product identification. This periodic processing maintains adequate recognition accuracy while significantly reducing power consumption compared to continuous processing
Solution Approach 2:
The system performs preliminary actions by pre-processing images to identify and isolate the area of interest before full product recognition. The computer hardware first processes the image to locate and segment the product area, then applies full recognition algorithms only to this isolated region. This preliminary action reduces the overall processing load and power consumption while maintaining high recognition accuracy for the actual product identification
3Productivity
If the entire field of view is processed for product identification, then no products are missed, but processing efficiency decreases
Solution Approach 1:
The patent segments the field of view into different regions, specifically identifying and isolating the area of interest where products are actually placed in the shopping cart. The computer hardware processes only this segmented area of interest rather than the entire field of view. This segmentation maintains product detection completeness within the relevant area while dramatically improving processing efficiency by excluding irrelevant background regions
Data Source
AI summary
A system and method for a low-power, self-checkout shopping device. A system and method of operations include one or more image sensors arranged on the shopping device, and computer hardware connected with the one or more image sensors. The computer hardware is configured to perform operations including demarcate an area of interest associated with the shopping receptacle within a field of view (FOV) of the one or more image sensors, and detect one or more features of one or more products that are imaged by the one or more image sensors in the area of interest associated with the shopping receptacle. The operations further include associate the detected features of the one or more products with a product database to determine a product identification for each of the one or more products that are imaged by the one or more image sensors.