Robust In-the-Wild Exercise Recognition from a Single Wearable: Data-Side Fusion, Sensor Rotation, and Feature Engineering

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Phan, Hoang Khang, Le, Khang, Nguyen, Tu Nhat Khang, Van Dao, Anh, Le, Nhat Tan
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866912802492383232
author Phan, Hoang Khang
Le, Khang
Nguyen, Tu Nhat Khang
Van Dao, Anh
Le, Nhat Tan
author_facet Phan, Hoang Khang
Le, Khang
Nguyen, Tu Nhat Khang
Van Dao, Anh
Le, Nhat Tan
contents Monitoring physical exercises is vital for health promotion, with automated systems becoming standard in personal health surveillance. However, sensor placement variability and unconstrained movements limit their effectiveness. This study proposes the team "3KA"'s one-sensor workout activity recognition method using feature extraction and data augmentation in 2ndWEAR Dataset Challenge. From raw acceleration, angle and signal magnitude vector features were derived, followed by extraction of statistical, fractal/spectral, and higher-order differential features. A fused dataset combining left/right limb data was created, and augmented via sensor rotation and axis inversion. We utilized a soft voting model combining Hist Gradient Boosting with balanced weights and Extreme Gradient Boosting without. Under group 5-fold evaluation, the model achieved 58.83\% macro F1 overall (61.72% arm, 55.95% leg). ANOVA F-score showed fractal/spectral features were most important for arm-based recognition but least for leg-based. The code to reproduce the experiments is publicly available via: https://github.com/Khanghcmut/WEAR\_3K
format Preprint
id arxiv_https___arxiv_org_abs_2511_23173
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Robust In-the-Wild Exercise Recognition from a Single Wearable: Data-Side Fusion, Sensor Rotation, and Feature Engineering
Phan, Hoang Khang
Le, Khang
Nguyen, Tu Nhat Khang
Van Dao, Anh
Le, Nhat Tan
Human-Computer Interaction
Monitoring physical exercises is vital for health promotion, with automated systems becoming standard in personal health surveillance. However, sensor placement variability and unconstrained movements limit their effectiveness. This study proposes the team "3KA"'s one-sensor workout activity recognition method using feature extraction and data augmentation in 2ndWEAR Dataset Challenge. From raw acceleration, angle and signal magnitude vector features were derived, followed by extraction of statistical, fractal/spectral, and higher-order differential features. A fused dataset combining left/right limb data was created, and augmented via sensor rotation and axis inversion. We utilized a soft voting model combining Hist Gradient Boosting with balanced weights and Extreme Gradient Boosting without. Under group 5-fold evaluation, the model achieved 58.83\% macro F1 overall (61.72% arm, 55.95% leg). ANOVA F-score showed fractal/spectral features were most important for arm-based recognition but least for leg-based. The code to reproduce the experiments is publicly available via: https://github.com/Khanghcmut/WEAR\_3K
title Robust In-the-Wild Exercise Recognition from a Single Wearable: Data-Side Fusion, Sensor Rotation, and Feature Engineering
topic Human-Computer Interaction
url https://arxiv.org/abs/2511.23173