A deep-learning based gait classification and anomaly detection framework for healthcare surveillance
- 1Department of Health Informatics, College of Applied Medical Sciences, Qassim University
Abstract
Gait classification and anomaly detection are non-intrusive approaches that can support healthcare surveillance and clinical gait analysis by identifying abnormal walking patterns. Despite recent advancements, existing methods remain limited by model complexity, high computational requirements, and privacy concerns. This study proposes a unified framework that combines three complementary components: (i) transformer-based temporal modeling to capture both short-term and long-term gait dynamics, (ii) lightweight architectures for efficient deployment on edge devices, and (iii) federated learning (FL) for privacy-preserving distributed training without requiring raw data sharing. Experiments were conducted on a balanced subset of 60,000 images from the Gait Detection Processed dataset using a controlled federated learning environment designed as a proof-of-concept evaluation rather than a large-scale deployment setting. The dataset consisted of three categories: background/non-gait, normal gait, and abnormal gait. Vision Transformer (ViT), ConvLSTM, and MobileViT architectures were evaluated for three-class gait classification and anomaly detection under a federated learning setting. Among the evaluated models, MobileViT-Large achieved the highest performance with 97.2% accuracy, 96.8% precision, 97.5% recall, and 97.1% F1-score, although it required higher computational resources and showed greater overfitting tendencies. MobileViT-Small achieved the best balance between efficiency and performance with 94.0% accuracy, making it more suitable for edge deployment. SHAP-based analysis further showed that the models focused on meaningful gait regions, such as torso and limb movements. This proposed framework provides a comparative benchmark of recurrent and transformer-based architectures within a privacy-preserving framework for healthcare monitoring applications.
Related
- Traffic surveillance video-based a unified road accident detection, classification, and vehicle tracking framework using an optimized spatio-temporal attention mechanism
- A deep learning-based silhouette gait detection framework for fall risk classification in Parkinson’s disease
- Hybrid deep learning–driven explainable AI framework for fault detection and classification in smart power grids
- A WHO CNS5-Aligned Classification Framework for Population-Based Surveillance of Brain and Other Central Nervous System Tumors in the United States
- A lightweight YOLO based framework for real-time seat belt detection in intelligent traffic surveillance
BibTeX
@article{alfridi2026deeplearning,
title = {{A deep-learning based gait classification and anomaly detection framework for healthcare surveillance}},
author = {Mohammad F. Alfridi},
journal = {Scientific Reports},
year = {2026},
month = jul,
doi = {10.1038/s41598-026-62788-6},
url = {https://doi.org/10.1038/s41598-026-62788-6},
note = {CC BY},
}
RIS
TY - JOUR
TI - A deep-learning based gait classification and anomaly detection framework for healthcare surveillance
AU - Mohammad F. Alfridi
PY - 2026
DA - 2026/07/01
JO - Scientific Reports
DO - 10.1038/s41598-026-62788-6
UR - https://doi.org/10.1038/s41598-026-62788-6
AB - Gait classification and anomaly detection are non-intrusive approaches that can support healthcare surveillance and clinical gait analysis by identifying abnormal walking patterns. Despite recent advancements, existing methods remain limited by model complexity, high computational requirements, and privacy concerns. This study proposes a unified framework that combines three complementary components: (i) transformer-based temporal modeling to capture both short-term and long-term gait dynamics, (ii) lightweight architectures for efficient deployment on edge devices, and (iii) federated learning (FL) for privacy-preserving distributed training without requiring raw data sharing. Experiments were conducted on a balanced subset of 60,000 images from the Gait Detection Processed dataset using a controlled federated learning environment designed as a proof-of-concept evaluation rather than a large-scale deployment setting. The dataset consisted of three categories: background/non-gait, normal gait, and abnormal gait. Vision Transformer (ViT), ConvLSTM, and MobileViT architectures were evaluated for three-class gait classification and anomaly detection under a federated learning setting. Among the evaluated models, MobileViT-Large achieved the highest performance with 97.2% accuracy, 96.8% precision, 97.5% recall, and 97.1% F1-score, although it required higher computational resources and showed greater overfitting tendencies. MobileViT-Small achieved the best balance between efficiency and performance with 94.0% accuracy, making it more suitable for edge deployment. SHAP-based analysis further showed that the models focused on meaningful gait regions, such as torso and limb movements. This proposed framework provides a comparative benchmark of recurrent and transformer-based architectures within a privacy-preserving framework for healthcare monitoring applications.
SN - 2045-2322
ER -