Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

HarMI: Human Activity Recognition Via Multi-Modality Incremental Learning.

Xiao Zhang, Hongzheng Yu, Yang Yang

    IEEE Journal of Biomedical and Health Informatics
    |June 1, 2021
    PubMed
    Summary
    This summary is machine-generated.

    Related Concept Videos

    You might also read

    Related Articles

    Articles linked to this work by shared authors, journal, and citation graph.

    Sort by
    Same author

    Tongfu Pingchuan decoction enhances Chrm4 expression in intestinal macrophages to alleviate sepsis in rats.

    3 Biotech·2026
    Same author

    MPQ-DMv2: Flexible Residual Mixed Precision Quantization for Low-Bit Diffusion Models With Temporal Distillation.

    IEEE transactions on pattern analysis and machine intelligence·2026
    Same author

    Pathological Complete Response in Advanced Hepatocellular Carcinoma With Portal Vein Tumor Thrombus Following Conversion Therapy With TACE, HAIC, Lenvatinib, and Sintilimab: A Case Report.

    Clinical case reports·2026
    Same author

    Deciphering the glioblastoma microenvironmental landscape with multi-modal radiogenomics to guide prognosis and personalized therapy.

    Journal of translational medicine·2026
    Same author

    All-trans retinoic acid combined with the VAH regimen for EVI1-positive acute myeloid leukemia: a case report with a brief literature review.

    Frontiers in oncology·2026
    Same author

    Interventions for enhancing resilience in breast cancer patients: a realist review.

    Supportive care in cancer : official journal of the Multinational Association of Supportive Care in Cancer·2026

    This study introduces HarMI, a novel multi-modality incremental learning model for human activity recognition (HAR). HarMI efficiently learns new activities without retraining, reducing storage and time costs.

    Area of Science:

    • Computer Science
    • Machine Learning
    • Signal Processing

    Background:

    • Human Activity Recognition (HAR) is crucial for healthcare and smart cities, utilizing smartphone and wearable sensors.
    • Current HAR methods often require offline training, demanding large storage and costly retraining for new activities.

    Purpose of the Study:

    • To develop an efficient multi-modality incremental learning model for HAR.
    • To enable continuous learning of new activities without storing prior data or extensive retraining.

    Main Methods:

    • Proposed HarMI model employing an attention mechanism for heterogeneous sensor data alignment.
    • Utilized elastic weight consolidation and canonical correlation analysis to mitigate catastrophic forgetting in incremental learning.

    Related Experiment Videos

    Main Results:

    • HarMI demonstrated superior performance compared to state-of-the-art methods in experiments.
    • The model achieved efficient training with minimal storage requirements.

    Conclusions:

    • HarMI offers a continuous learning solution for sensor-based HAR, overcoming limitations of offline methods.
    • The proposed approach facilitates rapid adaptation to new activities with reduced computational overhead.