一种基于少量标签数据的多阶段人体活动识别方法

By combining the SimCLR framework and the FixMatch model, human activity recognition is achieved using a small amount of labeled data, which solves the problem of deep learning's dependence on a large amount of labeled data, realizes efficient multi-stage human activity recognition, and improves the model's recognition performance and generalization ability.

CN117932434BActive Publication Date: 2026-07-17HARBIN INST OF TECH
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
HARBIN INST OF TECH
Filing Date
2024-01-25
Publication Date
2026-07-17

Smart Images

  • Figure CN117932434B_ABST
    Figure CN117932434B_ABST
Patent Text Reader

Abstract

本发明公开了一种基于少量标签数据的多阶段人体活动识别方法，所属技术领域为智能物联网、普适计算领域，包括：获取IMU数据，对IMU数据进行预处理，获得预处理数据；基于SimCLR框架和预处理数据进行对比学习，获得特征提取器；通过特征提取器对预处理数据进行处理，生成有标签数据的表征，通过有标签数据的表征进行监督训练，获得特定于人体活动识别的模型；将训练后的FixMatch模型和特定于人体活动识别的模型进行结合，获得人体活动识别模型，基于人体活动识别模型对多阶段人体活动进行识别。本发明将对比学习和半监督学习相结合，充分利用大量的未标记IMU数据，使得活动识别模型仍然可以具有令人满意的识别效果。
Need to check novelty before this filing date? Find Prior Art