Lost article retrieving system and method

By introducing convolutional neural networks with attention mechanisms and hierarchical retrieval strategies, combined with offline verification mechanisms, the problems of inaccurate feature extraction, disconnect between online and offline processes, and insufficient privacy protection in existing lost and found systems have been solved, achieving efficient and accurate lost and found retrieval and cost optimization.

CN122368544APending Publication Date: 2026-07-10TIANJIN HAIWEI TRANSPORTATION FACILITIES TECHNOLOGY DEVELOPMENT CO LTD
0 Cites 0 Cited by

Patent Information

Application Number
CN202610505478.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-04-16
Publication Date
2026-07-10

AI Technical Summary

Technical Problem

The existing lost and found management system suffers from inaccurate feature extraction, lack of online and offline verification loop, imperfect incentive mechanism, and insufficient privacy protection, resulting in low accuracy, poor efficiency, and high cost in finding lost items.

Method used

Feature extraction is performed using a convolutional neural network with an attention mechanism. Combined with a hierarchical retrieval strategy and an offline verification mechanism, a closed-loop service is built through the franchisee's terminal to achieve accurate matching of items and online-offline collaboration. Item feature vectors are used to replace the original image storage, and the reward mechanism is standardized.

Benefits of technology

It improves the accuracy and efficiency of lost item retrieval, reduces server load and manual review costs, protects user privacy, and achieves a complete service loop and a reasonable incentive mechanism.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention relates to the technical field of data processing and Internet of Things, and particularly discloses a lost article retrieving system and method. A user side collects and uploads an article image; the server extracts article feature vectors through a convolutional neural network containing an attention mechanism, performs matching by adopting a hierarchical strategy of'type coarse screening + feature fine screening ', and realizes reward freezing issuing and service process recording in combination with a reward management module and an order management module; and the offline server is responsible for article verification. According to the method, key features of the article can be focused, the comparison complexity is reduced, an online and offline complete service closed loop is constructed, false uploading is prevented, the privacy of the user is protected, and the expandability of the modular design is high.
Need to check novelty before this filing date? Find Prior Art