Q-PPG: Energy-Efficient PPG-Based Heart Rate Monitoring on Wearable Devices (2021-10-21T00:00:00.000000Z)

TL;DR

A design space exploration methodology to automatically generate a rich family of deep Temporal Convolutional Networks (TCNs) for HR monitoring, all derived from a single “seed” model, whose most accurate model sets a new state-of-the-art in Mean Absolute Error.

Abstract

Hearth Rate (HR) monitoring is increasingly performed in wrist-worn devices using low-cost photoplethysmography (PPG) sensors. However, Motion Artifacts (MAs) caused by movements of the subject’s arm affect the performance of PPG-based HR tracking. This is typically addressed coupling the PPG signal with acceleration measurements from an inertial sensor. Unfortunately, most standard approaches of this kind rely on hand-tuned parameters, which impair their generalization capabilities and their applicability to real data in the field. In contrast, methods based on deep learning, despite their better generalization, are considered to be too complex to deploy on wearable devices. In this work, we tackle these limitations, proposing a design space exploration methodology to automatically generate a rich family of deep Temporal Convolutional Networks (TCNs) for HR monitoring, all derived from a single “seed” model. Our flow involves a cascade of two Neural Architecture Search (NAS) tools and a hardware-friendly quantizer, whose combination yields both highly accurate and extremely lightweight models. When tested on the PPG-Dalia dataset, our most accurate model sets a new state-of-the-art in Mean Absolute Error. Furthermore, we deploy our TCNs on an embedded platform featuring a STM32WB55 microcontroller, demonstrating their suitability for real-time execution. Our most accurate quantized network achieves 4.41 Beats Per Minute (BPM) of Mean Absolute Error (MAE), with an energy consumption of 47.65 mJ and a memory footprint of 412 kB. At the same time, the smallest network that obtains a MAE $< $ 8 BPM, among those generated by our flow, has a memory footprint of 1.9 kB and consumes just 1.79 mJ per inference.

References58 items

Pruning In Time (PIT): A Lightweight Network Architecture Optimizer for Temporal Convolutional Networks

Cloud–Edge-Based Lightweight Temporal Convolutional Networks for Remaining Useful Life Prediction in IIoT

NAS-PPG: PPG-Based Heart Rate Estimation Using Neural Architecture Search

Robust and Energy-Efficient PPG-Based Heart-Rate Monitoring

H-Watch: An Open, Connected Platform for AI-Enhanced COVID19 Infection Symptoms Monitoring and Contact Tracing.

An Ensemble of Hyperdimensional Classifiers: Hardware-Friendly Short-Latency Seizure Detection With Automatic iEEG Electrode Selection

DORY: Automatic End-to-End Deployment of Real-World DNNs on Low-Cost IoT MCUs

MCUNet: Tiny Deep Learning on IoT Devices

Robust PPG-based Ambulatory Heart Rate Tracking Algorithm

Heart Rate Monitoring using Sparse Spectral Curve Tracing

Binary CorNET: Accelerator for HR Estimation From Wrist-PPG

Rethinking Differentiable Search for Mixed-Precision Neural Networks

FBNetV2: Differentiable Neural Architecture Search for Spatial and Channel Dimensions

CMix-NN: Mixed Low-Precision CNN Library for Memory-Constrained Edge Devices

FitBit

Robust Real-Time Embedded EMG Recognition Framework Using Temporal Convolutional Networks on a Multicore IoT Processor

Deep PPG: Large-Scale Heart Rate Estimation with Convolutional Neural Networks

Finite State Machine Framework for Instantaneous Heart Rate Validation Using Wearable Photoplethysmography During Intensive Exercise

A Survey on Deep Learning based Brain Computer Interface: Recent Advances and New Frontiers

Evaluation of Coherence Between ECG and PPG Derived Parameters on Heart Rate Variability and Respiration in Healthy Volunteers With/Without Controlled Breathing

CorNET: Deep Learning Framework for PPG-Based Heart Rate Estimation and Biometric Identification in Ambulant Environment

A Review of Wrist-Worn Wearable: Sensors, Models, and Challenges

ProxylessNAS: Direct Neural Architecture Search on Target Task and Hardware

A review on wearable photoplethysmography sensors and their potential future applications in health care

MnasNet: Platform-Aware Neural Architecture Search for Mobile

An Empirical Evaluation of Generic Convolutional and Recurrent Networks for Sequence Modeling

PACT: Parameterized Clipping Activation for Quantized Neural Networks

MorphNet: Fast & Simple Resource-Constrained Structure Learning of Deep Networks

Computationally efficient heart rate estimation during physical exercise using photoplethysmographic signals

Accurate Heart Rate Monitoring During Physical Exercises Using PPG

Evaluating the accuracy of wearable heart rate monitors

Temporal Convolutional Networks: A Unified Approach to Action Segmentation

A Novel Time-Varying Spectral Filtering Algorithm for Reconstruction of Motion Artifact Corrupted Heart Rate Signals During Intense Physical Activities Using a Wearable Photoplethysmogram Sensor

Deep Compression: Compressing Deep Neural Network with Pruning, Trained Quantization and Huffman Coding

Human photoplethysmogram: new insight into chaotic characteristics

Heart Rate Tracking using Wrist-Type Photoplethysmographic (PPG) Signals during Physical Exercise with Simultaneous Accelerometry

Photoplethysmography-Based Heart Rate Monitoring in Physical Activities via Joint Sparse Spectrum Reconstruction

Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift

TROIKA: A General Framework for Heart Rate Monitoring Using Wrist-Type Photoplethysmographic Signals During Intensive Physical Exercise

Wearable Photoplethysmographic Sensors—Past and Present

Neural Networks: Tricks of the Trade

A and V

“Neurosky,”

and s

Hyperdimensional Computing with Local Binary Patterns: One-shot Learning for Seizure Onset Detection and Identification of Ictogenic Brain Regions from Short-time iEEG Recordings

Empatica

I and J

Γ and B

How to Train Neural Networks

Heart Rate Tracking using Wrist-Type Photoplethysmographic ( PPG ) Signals during Physical Exercise with Simultaneous Accelerometry

and as an in

NXP. Kv4xp100m168

Respiban professional 2019

are with the Department of Politecnico di Torino,Turin, Italy

is also with the Department of Information Technology and Electrical Engineering at the ETH Zurich, 8092 Zurich, Switzerland

STMicroelectronics

Apple watch series

Reggio Emilia, Italy