A target event logical judgment method and system based on visual perception and knowledge graph reverse optimization

By constructing an RDF triple knowledge graph of entity-relationship-attribute and a logical decision function, and optimizing parameters by combining labeled data, the closed-loop optimization problem of the visual data logical decision system is solved, achieving efficient and interpretable event determination, which is applicable to fields such as industrial quality inspection and intelligent security.

CN122452803APending Publication Date: 2026-07-24NANJING COMPREHENSIVE SAFETY CONSULTING CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
NANJING COMPREHENSIVE SAFETY CONSULTING CO LTD
Filing Date
2026-05-07
Publication Date
2026-07-24

Smart Images

  • Figure QLYQS_3
    Figure QLYQS_3
  • Figure QLYQS_4
    Figure QLYQS_4
Patent Text Reader

Abstract

The application provides a target event logical judgment method and system based on visual perception and knowledge graph reverse optimization, comprising the following steps: S1, identifying target entities and visual features of scene images or video data; S2, defining a structured fact set; S3, constructing a knowledge graph; S4, determining a flow to obtain a judgment score S of an event j ; S5, obtaining optimized parameters; S6, updating the parameters to obtain a judgment score of the event after the parameters are optimized. The application also provides a target event logical judgment system based on visual perception and knowledge graph reverse optimization. The technical scheme can quantify the contribution degree of different precondition to event judgment, avoid one-sidedness of single condition judgment, further reduce judgment deviation through real labeled data iterative optimization parameters, make event judgment result more close to real scene, effectively improve the accuracy of event recognition in complex scene, and has high explainability and flexibility.
Need to check novelty before this filing date? Find Prior Art