EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Deep learning for stress oriented human activity recognition

Muhammad Hamza, Nasir Uddin, Gulnaz Anjum, Mohammad Anas, Uzair Gabol, Nida Saddaf Khan

Frontiers in Digital Health · 18.05.2026 · 8

DOI: 10.3389/fdgth.2026.1729804 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Frontiers in Digital Health, online erschienen am 18.05.2026. Verfasst von Muhammad Hamza und 5 weiteren Autor:innen. Beteiligte Einrichtung: Department of Computer Science, National University of Computer and Emerging Sciences, Karachi, Pakistan. Die Arbeit nennt 29 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Universitetet i Oslo.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Introduction

Human Activity Recognition (HAR) using sensor-generated time-series data has gained significant attention for assessing mental and physical states to address various behavioral disorders. This study utilizes benchmark datasets of stress-related activities to improve prediction performance.

Methods

To achieve this, we employ multiple deep learning architectures, including Recurrent Neural Networks (RNNs), Long Short-Term Memory (LSTM) networks, and Transformer models, for feature extraction and classification. Comprehensive experiments are conducted to evaluate model performance, with particular focus on the impact of window size and overlap ratio on classification accuracy.

Results

The experimental results demonstrate that Transformer models outperform LSTM and RNN models, achieving classification accuracies of 97.83%, 97.36%, and 92.4% on the test dataset, respectively. Furthermore, the proposed approach shows a significant improvement over the deep neural network reported in the original Stressense dataset study.

Discussion

These findings highlight the effectiveness of Transformer-based architectures for HAR tasks involving stress detection. The improvement in classification performance suggests strong potential for advancing seamless mental health monitoring using non-intrusive wearable devices.

Schlagwörter

Stichwörter der Autor:innen

human activity recognition (HAR)long short-term memory (LSTM)mental healthrecurrent neural networks (RNNs)stress detectiontransformers

Autor:innen und Institutionen

  • Department of Computer Science, National University of Computer and Emerging Sciences, Karachi, Pakistan.
  • Department of Psychology, University of Oslo, Oslo, Norway.
  • University of Texas Health Science Center, Houston, TX, United States.

Förderung

  • Universitetet i Oslo

Zitieren

APA

Hamza, M., Uddin, N., Anjum, G., Anas, M., Gabol, U., & Saddaf Khan, N. (2026). Deep learning for stress oriented human activity recognition. Frontiers in Digital Health, 8. https://doi.org/10.3389/fdgth.2026.1729804

Vancouver

Hamza M, Uddin N, Anjum G, Anas M, Gabol U, Saddaf Khan N. Deep learning for stress oriented human activity recognition. Frontiers in Digital Health. 2026;8. doi:10.3389/fdgth.2026.1729804

BibTeX
@article{Hamza2026,
  author = {Hamza, Muhammad and Uddin, Nasir and Anjum, Gulnaz and Anas, Mohammad and Gabol, Uzair and Saddaf Khan, Nida},
  title = {Deep learning for stress oriented human activity recognition},
  journal = {Frontiers in Digital Health},
  year = {2026},
  volume = {8},
  doi = {10.3389/fdgth.2026.1729804}
}
Kostenlos starten →