Computer Vision Self-Checkout Guidance for Item Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Customers face challenges in accurately completing transactions at self-checkout lanes due to issues like overlapping items, occlusion, and poor visibility, which affect the accuracy of computer vision systems, leading to potential errors and increased computational resources.
Innovation Solution
Implementing a computer vision system that predicts checkout issues and provides real-time instructions to customers using ML models, such as a computer vision ML model and additional sensor data, to address visibility and recognition problems, reducing computational strain and improving transaction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If computer vision systems are used to identify items at self-checkout lanes, then automation and productivity are improved, but measurement precision deteriorates due to overlapping items, occlusion, and poor visibility
Solution Approach 1:
The system provides real-time feedback to customers through the user interface when items are not properly positioned, explaining why recognition failed and guiding customers to adjust item placement. This feedback loop enables customers to correct positioning issues, allowing the automated computer vision system to achieve accurate item recognition without requiring manual intervention.
Solution Approach 2:
The system empowers customers to self-correct item positioning issues by providing them with actionable guidance through the user interface. Instead of requiring employee assistance or complex system interventions, customers can independently adjust item placement based on system feedback, maintaining automation while improving recognition accuracy.
2Loss of information
If computer vision systems process all item data without guidance, then complete data capture is achieved, but loss of time increases due to repeated processing attempts
Solution Approach 1:
The system performs preliminary analysis of captured images to determine whether items are properly positioned for recognition before initiating full processing. When items are improperly positioned, the system provides guidance to customers to correct the issue before final recognition attempts, preventing repeated processing cycles and reducing transaction time while maintaining complete data capture.
3Ease of operation
If detailed instructions are provided to customers, then ease of operation improves, but device complexity increases
Solution Approach 1:
The user interface acts as an intermediary that translates complex computer vision analysis into simple, actionable instructions for customers. Rather than requiring the checkout system itself to be complex, the UI mediates between the sophisticated image processing and the customer, presenting information in an easily understandable format that improves operation without significantly increasing system complexity.
Data Source
AI summary
Techniques relating to using machine learning (ML) with a point of sale (POS) system. These techniques include identifying one or more images, captured at a POS system, of one or more items for purchase. The techniques further include predicting a checkout issue relating to the one or more items for purchase, including determining the checkout issue using a trained ML model, based on the one or more images. The techniques further include generating one or more instructions for a purchaser of the one or more items, based on the predicted checkout issue, and presenting the one or more instructions at a user interface of the POS system.