Automated Scan-Zone Image Capture for Vision Checkout Training
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The manual process of capturing images of items in various positions within a scan zone for machine learning models in vision self-checkout is tedious and time-consuming, requiring significant human effort and time, especially when considering the large number of items in a store and the need for multiple images per item from different angles.
Innovation Solution
A system and method for automated item image capture and registration using a spiral component that rotates and moves items around X-Y coordinates within a scan zone, with multiple cameras capturing images from different angles and positions, while maintaining metadata for each image, reducing human intervention.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual image capture process is used with human trainer physically moving items, then image quality and variety can be ensured, but time consumption and labor effort increase significantly
Solution Approach 1:
The system enables self-service automated image capture where the spiral component autonomously moves items through the scan zone and triggers camera captures without human intervention. The processor automatically coordinates camera activation at multiple X-Y coordinates, eliminating the need for human trainers to manually position items while maintaining comprehensive image collection.
Solution Approach 2:
The patent replaces the manual mechanical system of human trainers physically moving items with an automated mechanical spiral component that rotates and translates items through the scan zone. This mechanical substitution enables precise, repeatable positioning at multiple coordinates while dramatically reducing labor requirements and time consumption.
2Reliability
If multiple images per item from various positions are captured, then machine learning model accuracy improves, but data collection time and complexity increase
Solution Approach 1:
The patent segments the image collection process into discrete X-Y coordinate positions within the scan zone. The spiral component stops at multiple predetermined positions, and cameras capture images at each segment. This segmentation enables systematic collection of diverse item views while simplifying the overall process through automated coordination of position and capture timing.
Solution Approach 2:
The spiral component performs multiple functions: it rotates items to present different faces, translates items to different X-Y coordinates, and triggers camera captures at each position. This multi-functionality enables comprehensive image collection from various angles and positions using a single integrated mechanism, reducing the need for multiple separate systems.
3Reliability
If comprehensive item coverage across hundreds of items is achieved, then vision checkout accuracy improves, but manual effort and time requirements become excessive
Solution Approach 1:
The spiral component enables continuous image capture operations by automatically rotating through multiple X-Y coordinates and triggering cameras at each position without manual intervention. This continuous automated operation allows rapid processing of hundreds of items, maintaining comprehensive image collection while dramatically increasing productivity compared to manual methods.
Data Source
AI summary
An apparatus is provided that executes instructions to move an item within a scan zone to specific X-Y coordinates of the scan zone. The item is placed on a platform and is rotated 360 degrees at each X-Y coordinate within the scan zone. Item images are captured by cameras at each X-Y coordinate and for each rotation at the corresponding X-Y coordinate. The item images are labeled and retained. The item images are used as input to a Machine-Learning Model (MLM) to train the MLM to recognize item codes for the items when subsequent images are captured for the item during a checkout. In an embodiment, during a checkout unknown item images are flagged and labeled with the corresponding item code when the corresponding item's barcode is scanned during the checkout; the labeled item images are also retailed for training the MLM for item recognition.