An attribute-driven dynamic model optimization method and system

By controlling the dynamic activation of attribute recognition and self-training, and combining attribute association rules and user corrections, the attribute recognition model is optimized, solving the problem of illogical recognition results in existing technologies, improving recognition accuracy and adaptability, and realizing the model's self-learning and optimization.

CN120976684BActive Publication Date: 2026-01-20SHENZHEN TIEYUE ELECTRIC CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202511501840.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2025-10-21
Publication Date
2026-01-20
Estimated Expiration
2045-10-21

AI Technical Summary

Technical Problem

Existing deep learning-based attribute recognition methods suffer from illogical and common-sense results in practical deployments, leading to invalid alarms and erroneous events, and low recognition accuracy.

Method used

By acquiring algorithm configuration parameters, the activation and deactivation of attribute recognition and self-training are controlled. Initial attribute information is obtained using object detection and attribute recognition models, and post-processing is performed based on preset attribute association rules to generate final attribute information. User corrections are received and model training is performed. Attribute alarm rules, coexistence rules, and mutual exclusion rules are introduced for refined processing.

Benefits of technology

It improves the accuracy and adaptability of attribute recognition, reduces false alarms, enhances the real-time performance and interactivity of the system, ensures high quality of labeled data, and enables the model to learn and optimize itself.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120976684B_ABST
    Figure CN120976684B_ABST
Patent Text Reader

Abstract

The application discloses a kind of attribute-driven dynamic model optimization method and system, it is related to data processing field.In the method, target detection model is used to process the image to be detected, and the target object in the image to be detected is obtained;If it is determined that attribute recognition switch is opened, the target object is input into attribute recognition model, and the initial attribute information corresponding to the target object is obtained;Based on the preset attribute association rule, the initial attribute information is post-processed to generate the final attribute information;When it is determined that the final attribute information meets the preset alarm condition, an alarm event is generated;In response to the alarm event, the target object and attribute information are displayed on the user interface, and the user's correction operation on the final attribute information is received to obtain labeled data;When it is determined that the self-training switch is triggered, the attribute recognition model is trained based on the labeled data to obtain the target attribute recognition model.Implementation of the technical solution provided in the application improves the accuracy of the model in identifying images.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Construction site personnel attribute detection method, system and device and medium

    CN114140745A

  • Interstitial tumor mitotic cell recognition and counting method based on deep learning

    CN118918096A