Unmanned Retail Product Delivery Recognition via Video Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In smart retail stores, existing systems struggle to accurately recognize product deliveries between customers, leading to discrepancies in product information, which affects inventory accuracy and efficiency.
Innovation Solution
A method and device that acquire videos from cameras, perform recognition on video segments to identify delivery initiation and reception users, and update product information using a preset delivery recognition model, ensuring accurate tracking of product transfers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If gravity sensors and cameras are installed on shelves to recognize products taken by customers, then product recognition capability is improved, but the system cannot recognize product deliveries between customers, leading to information inaccuracy
Solution Approach 1:
The patent combines multiple detection technologies (gravity sensors, cameras, and delivery recognition models) into an integrated system. The gravity sensor data, video footage, and delivery recognition model work together to comprehensively track product movements, including both product removals and deliveries between customers, thereby resolving the information loss problem while maintaining recognition accuracy
Solution Approach 2:
The patent introduces a delivery recognition model as an intermediary component that processes video segments and gravity sensor data to specifically identify product delivery actions between customers. This intermediary layer bridges the gap between raw sensor data and accurate product delivery recognition, enabling the system to capture information that would otherwise be lost
2Measurement precision
If video recognition is performed on all videos to identify product deliveries, then delivery recognition accuracy is improved, but system complexity and processing time increase
Solution Approach 1:
The patent segments video data into specific segments that contain potential delivery actions, rather than processing entire videos continuously. By dividing the video stream into manageable segments and applying recognition models only to relevant portions, the system maintains high delivery recognition accuracy while significantly reducing computational complexity and processing requirements
3Loss of information
If comprehensive video recognition is performed to track all product movements, then product information accuracy is improved, but processing efficiency decreases
Solution Approach 1:
The patent performs preliminary analysis on video segments to identify potential delivery actions before applying the full delivery recognition model. This preliminary filtering step prepares data in advance by identifying regions of interest and pre-processing video segments, ensuring that comprehensive product information tracking is achieved while maintaining high processing efficiency through staged analysis
Data Source
AI summary
The present disclosure provides a method and a device for recognizing a product, an electronic device and a non-transitory computer readable storage medium, relating to a field of unmanned retail product recognition. The method includes the following. A video taken by each camera in a store is acquired. A recognition is performed on each video to obtain a video segment that a product delivery is recognized and to obtain participated users. The participated users include a delivery initiation user and a delivery reception user. The video segment is inputted into a preset delivery recognition model to obtain a recognition result. The recognition result includes a product delivered and a delivery probability. The product information of products carried by the participated users is updated based on the recognition result.