A 1D-CNN Based Deep Learning Technique for Sleep Apnea Detection in IoT Sensors

TL;DR

This paper presents a high-resolution sleep apnea detection method using a 1D-CNN on ECG signals from wearable IoT devices, achieving high accuracy and exploring model optimization for resource-limited devices.

Contribution

Introduces a novel 1D-CNN based approach for second-by-second sleep apnea detection with high accuracy and analyzes model pruning and personalization for IoT wearables.

Findings

Achieves 99.56% accuracy and 96.05% sensitivity in apnea detection.

Pruned model with 80% sparsity maintains 97.34% accuracy.

Patient-specific models show high accuracy and sensitivity.

Abstract

Internet of Things (IoT) enabled wearable sensors for health monitoring are widely used to reduce the cost of personal healthcare and improve quality of life. The sleep apnea-hypopnea syndrome, characterized by the abnormal reduction or pause in breathing, greatly affects the quality of sleep of an individual. This paper introduces a novel method for apnea detection (pause in breathing) from electrocardiogram (ECG) signals obtained from wearable devices. The novelty stems from the high resolution of apnea detection on a second-by-second basis, and this is achieved using a 1-dimensional convolutional neural network for feature extraction and detection of sleep apnea events. The proposed method exhibits an accuracy of 99.56% and a sensitivity of 96.05%. This model outperforms several lower resolution state-of-the-art apnea detection methods. The complexity of the proposed model is analyzed. We also analyze the feasibility of model pruning and binarization to reduce the resource requirements on a wearable IoT device. The pruned model with 80\% sparsity exhibited an accuracy of 97.34% and a sensitivity of 86.48%. The binarized model exhibited an accuracy of 75.59% and sensitivity of 63.23%. The performance of low complexity patient-specific models derived from the generic model is also studied to analyze the feasibility of retraining existing models to fit patient-specific requirements. The patient-specific models on average exhibited an accuracy of 97.79% and sensitivity of 92.23%. The source code for this work is made publicly available.