Invention Title:

Distributed and Automated Transaction Systems

Publication number:

US20250037177

Publication date:
Section:

Physics

Class:

G06Q30/06

Inventors:

Applicant:

Drawings (4 of 14)

Smart overview of the Invention

The patent application describes systems and methods utilizing machine vision and machine learning to streamline retail transactions by minimizing user involvement. The primary concept involves a user simply removing an item from a retail location, which triggers the transaction process. The system is designed to enhance convenience, addressing common frustrations such as long checkout lines in traditional brick-and-mortar stores.

System Components

The system comprises a product enclosure with shelves, cameras, processors, and a network connection. The cameras capture images of the transaction, which are processed by the processors. These images are then transmitted to a remote back-end infrastructure for further analysis. The infrastructure identifies actions such as a hand reaching into the enclosure and determines the specific product being taken.

User Identification and Transaction Processing

User identification can be achieved through biometric data, machine-readable codes, or device-transmitted codes. Once a product is identified, the system charges the user's account for the transaction amount corresponding to the product. This process is automated to ensure seamless user experience without manual checkout steps.

Image Processing Techniques

The system employs various image processing techniques to optimize transaction accuracy. It discards irrelevant image subsets and uses localization models to focus on significant image portions. Additionally, it monitors changes within the product enclosure to identify interactions accurately and request necessary data uploads when changes are detected.

Security Features

Security is managed through controlled access to the product enclosure. The system can unlock doors in response to specific messages from the back-end infrastructure and monitors door status to ensure secure operations. Video recording and transmission protocols are also in place to document transactions and detect unauthorized access or prolonged door openings.