RFID-Video Fusion for Retail User Action Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current applications of image processing and RFID technologies in retail are limited to simple scenarios, such as video monitoring and inventory checks, failing to effectively determine user actions and item interests in offline retail environments.
Innovation Solution
A computer-implemented method and system that combines video processing and RFID technologies to recognize item displacement actions by users, determining the time and location of these actions and identifying target items in a non-stationary state through a classification model and RF signal analysis, thereby accurately determining user interest in specific items.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If image processing technology is applied for video monitoring, then user actions can be captured, but the system cannot accurately determine user interest in specific items
Solution Approach 1:
The patent combines image processing technology with RFID technology to create an integrated system. The image processing component captures user actions (flipping, moving items) while the RFID component identifies specific items through tag signals. By merging these two technologies, the system achieves both action recognition and item identification, resolving the limitation of using image processing alone.
Solution Approach 2:
The patent introduces RFID tags as an intermediary element attached to items. These tags serve as mediators between the user action detection system and the item identification system. When a user interacts with an item, the RFID tag detects the change in signal strength or distance, providing precise item identification that complements the visual action recognition.
2Measurement precision
If RFID technology is applied for inventory check, then item identification is achieved, but the system cannot recognize user actions
Solution Approach 1:
The patent merges RFID technology for item identification with image processing technology for action recognition. The RFID system continues to provide accurate item identification through tag signals, while the added image processing component captures and analyzes user actions (flipping, moving items). This combination ensures both item identification accuracy and user action information are captured simultaneously.
Solution Approach 2:
The integrated system performs multiple functions: it identifies items using RFID tags, recognizes user actions through image processing, determines user interest by correlating actions with items, and provides comprehensive retail analytics. This multi-functional approach eliminates the need for separate systems for inventory management and customer behavior analysis.
3Device complexity
If simple video monitoring is used, then the system structure remains simple, but it cannot determine user interest in specific items
Solution Approach 1:
The patent segments the system into distinct functional modules: an image processing module for capturing and analyzing user actions, an RFID module for item identification, and a data processing module for correlating actions with items to determine user interest. This segmentation allows the system to maintain modularity and relative simplicity while achieving advanced functionality through the coordinated operation of specialized components.
4Productivity
If code scanning for inventory check is applied, then inventory management is achieved, but user behavior analysis is not possible
Solution Approach 1:
The integrated system maintains efficient inventory management through RFID technology while simultaneously enabling user behavior analysis. The same RFID infrastructure that supports inventory tracking also provides item identification for action analysis, creating a universal platform that serves multiple retail functions without requiring separate systems.
Data Source
AI summary
The specification discloses a computer-implemented method for user action determination, comprising: recognizing an item displacement action performed by a user; determining a first time and a first location of the item displacement action; recognizing a target item in a non-stationary state; determining a second time when the target item is in the non-stationary state and a second location where the target item is in the non-stationary state; and in response to determining that the first time matches the second time and the first location matches the second location, determining that the item displacement action of the user is performed with respect to the target item.