CrossHAR: Generalizing Cross-dataset Human Activity Recognition via Hierarchical Self-Supervised Pretraining
Authors
ZH
Zhiqing Hong
ZL
Zelong Li
SZ
Shuxin Zhong
WL
Wenjun Lyu
HW
Haotian Wang
YD
Yi Ding
TH
Tian He
DZ
Desheng Zhang
Human Pose & Activity RecognitionPhysicians, Nurses & CliniciansPhysical Therapists & Rehabilitation SpecialistsAssistive Technology SpecialistsHCI Researchers
No results found
Research Questions / Practical Problems
Question signals indexed for this paper.
help
Research Questions
3- How can domain generalization for sensor-based human activity recognition be achieved on unknown target datasets?Category: HAR Model Robustness, Generalization, and Data AugmentationSimilar questionsarrow_forward
- How do physics-informed data augmentation and hierarchical self-supervised learning improve generalization of IMU (inertial measurement unit) models?Category: HAR Model Robustness, Generalization, and Data AugmentationSimilar questionsarrow_forward
- How can unlabeled IMU data best address domain shift problems?Category: HAR Model Robustness, Generalization, and Data AugmentationSimilar questionsarrow_forward
lightbulb
Practical Problems
1- Human activity recognition models in smart devices often fail in new scenarios.Category: HAR Model Robustness, Generalization, and Data AugmentationSimilar questionsarrow_forward
Based on Jaccard similarity of research subtopics & professions (≥60%)
Quick Actions
At a Glance
fact_checkPaper Snapshot
dataset
Source
UbiComp
calendar_month
Year
2024
emoji_events
Award
No award tagged
group
Authors
8 authors
sell
Subtopics
Human Pose & Activity Recognition
work
Professions
Physicians, Nurses & Clinicians, Physical Therapists & Rehabilitation Specialists, Assistive Technology Specialists, HCI Researchers
article
Content Status
Abstract only
hub
Related Papers
1 related papers