Sleep Position Classification using Transfer Learning for Bed-based Pressure Sensors

2025-05-12

AI Summary

To address the challenge of training deep learning models under clinical data scarcity, this study proposes a non-intrusive, four-class sleep posture classification method leveraging bed-based pressure-sensing mats (PSMs). Methodologically, it innovatively adapts vision foundation models—specifically ViTMAE for self-supervised pretraining and ViTPose for pose-aware representation learning—to low-resolution, time-series pressure image sequences, integrating temporal feature extraction with few-shot transfer learning. Evaluated on 112 real patient nights, the approach significantly outperforms baseline methods (TCN, SVM, XGBoost) in both classification accuracy and robustness. Furthermore, it demonstrates strong generalization capability on an independent, high-resolution dataset comprising 13 subjects. The resulting lightweight, computationally efficient framework is clinically deployable and supports scalable sleep quality monitoring and sleep disorder screening.

Technology Category

Computer Vision: Multi-modal VisionMachine Learning: Deep Neural Architectures and Foundation ModelsIntelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

Web Mining and Content Analysis: Large pretrained models with web dataSecurity and Privacy: Data transparency and provenanceGraph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Bed-based pressure-sensitive mats (PSMs) offer a non-intrusive way of monitoring patients during sleep. We focus on four-way sleep position classification using data collected from a PSM placed under a mattress in a sleep clinic. Sleep positions can affect sleep quality and the prevalence of sleep disorders, such as apnea. Measurements were performed on patients with suspected sleep disorders referred for assessments at a sleep clinic. Training deep learning models can be challenging in clinical settings due to the need for large amounts of labeled data. To overcome the shortage of labeled training data, we utilize transfer learning to adapt pre-trained deep learning models to accurately estimate sleep positions from a low-resolution PSM dataset collected in a polysomnography sleep lab. Our approach leverages Vision Transformer models pre-trained on ImageNet using masked autoencoding (ViTMAE) and a pre-trained model for human pose estimation (ViTPose). These approaches outperform previous work from PSM-based sleep pose classification using deep learning (TCN) as well as traditional machine learning models (SVM, XGBoost, Random Forest) that use engineered features. We evaluate the performance of sleep position classification from 112 nights of patient recordings and validate it on a higher resolution 13-patient dataset. Despite the challenges of differentiating between sleep positions from low-resolution PSM data, our approach shows promise for real-world deployment in clinical settings

Problem

• Classify sleep positions using bed-based pressure sensors
• Overcome limited labeled data with transfer learning
• Improve accuracy in low-resolution pressure-sensitive mat data

Innovation

• Transfer learning for sleep position classification
• Pre-trained Vision Transformer models adaptation
• Outperforms traditional machine learning methods