New AI model enhances Amazon checkout-free shopping

This model uses advanced AI techniques, including transformer-based machine learning, to simultaneously analyze data from cameras and sensors throughout a store.

A recent advancement in Amazon’s Just Walk Out technology introduces a new multi-modal foundation model designed to enhance the accuracy of checkout-free shopping. This model uses advanced AI techniques, including transformer-based machine learning, to simultaneously analyze data from cameras and sensors throughout a store.

This approach improves the system’s ability to accurately identify purchased items, even under challenging conditions like poor lighting or camera obstructions.

Just Walk Out technology utilizes a combination of cameras, weight sensors, and AI algorithms to determine what items are taken without requiring manual checkout.

Initially launched in 2018, Just Walk Out technology has relied on AI to monitor shopper behavior and track purchases. The new model builds on this by processing all sensor data in real-time, which allows it to adapt to various shopping scenarios more effectively.

It is trained using a 3D map of each store and an image catalog of merchandise, enabling it to recognize products even if store layouts change. This ensures accurate tracking of items and quantities, regardless of how stores are arranged, according to the company.

Related contents

Discuss