Skip to main content
QUICK REVIEW

[论文解读] Optimised Convolutional Neural Networks for Heart Rate Estimation and Human Activity Recognition in Wrist Worn Sensing Applications

Eoin Brophy, Willie Muehlhausen|arXiv (Cornell University)|Mar 30, 2020
Non-Invasive Vital Sign Monitoring参考文献 16被引用 12
一句话总结

本文提出一种仅使用单个光学传感器、低采样率的方法,通过优化的卷积神经网络(CNN)实现可穿戴设备中同时的心率估计与人体活动识别(HAR)。通过利用迁移学习进行HAR,以及使用CNN回归进行心率估计,该方法在5–10 Hz的采样率下实现了高精度,与传统256 Hz方法相比性能相当或更优,显著降低了功耗与成本,且未牺牲性能。

ABSTRACT

Wrist-worn smart devices are providing increased insights into human health, behaviour and performance through sophisticated analytics. However, battery life, device cost and sensor performance in the face of movement-related artefact present challenges which must be further addressed to see effective applications and wider adoption through commoditisation of the technology. We address these challenges by demonstrating, through using a simple optical measurement, photoplethysmography (PPG) used conventionally for heart rate detection in wrist-worn sensors, that we can provide improved heart rate and human activity recognition (HAR) simultaneously at low sample rates, without an inertial measurement unit. This simplifies hardware design and reduces costs and power budgets. We apply two deep learning pipelines, one for human activity recognition and one for heart rate estimation. HAR is achieved through the application of a visual classification approach, capable of robust performance at low sample rates. Here, transfer learning is leveraged to retrain a convolutional neural network (CNN) to distinguish characteristics of the PPG during different human activities. For heart rate estimation we use a CNN adopted for regression which maps noisy optical signals to heart rate estimates. In both cases, comparisons are made with leading conventional approaches. Our results demonstrate a low sampling frequency can achieve good performance without significant degradation of accuracy. 5 Hz and 10 Hz were shown to have 80.2% and 83.0% classification accuracy for HAR respectively. These same sampling frequencies also yielded a robust heart rate estimation which was comparative with that achieved at the more energy-intensive rate of 256 Hz.

研究动机与目标

  • 通过降低采样频率并消除对惯性传感器的需求,减少可穿戴设备的功耗与硬件复杂度。
  • 利用原始PPG信号的深度学习方法,提高运动伪影存在下的心率估计精度。
  • 证明仅使用光学PPG信号即可实现人体活动识别,无需额外传感器或高采样率。
  • 评估深度学习流水线在低采样频率(5–30 Hz)下的性能,与传统信号处理方法进行对比。
  • 优化基于CNN的回归模型,以实现从噪声PPG数据中稳健的心率估计。

提出的方法

  • 基于迁移学习的CNN在PPG信号上进行微调,以识别人体活动,将运动过程中信号形态的变化作为判别性特征。
  • 独立训练一个CNN回归模型(CNNR),直接将原始、噪声PPG信号映射到心率值,跳过传统滤波步骤。
  • 通过学习率、训练轮数和训练/测试集划分的超参数搜索,对CNNR进行优化,以最小化心率估计误差。
  • 评估5 Hz、10 Hz、15 Hz、30 Hz和256 Hz的采样频率,以分析低频下的性能退化情况。
  • 以开源的HeartPy工具包为基准,该工具使用传统信号处理方法进行心率估计。
  • 该方法完全依赖于光电容积脉搏波(PPG)数据,无需加速度计或多传感器融合。

实验结果

研究问题

  • RQ1在低采样率下,仅使用PPG信号是否能准确实现人体活动识别,而无需惯性传感器?
  • RQ2在运动伪影条件下,基于CNN的回归模型在心率估计方面是否优于传统信号处理方法(如HeartPy)?
  • RQ3将采样频率从256 Hz降低到5 Hz,对HAR和心率估计的精度有何影响?
  • RQ4一个单一的、优化的CNN模型是否能在多种活动和采样率下实现稳健性能?
  • RQ5何种超参数配置能实现噪声PPG信号中文心率估计误差最低?

主要发现

  • 在10 Hz采样频率下,HAR模型实现了83.0%的分类准确率,表明在低数据率下仍具有优异性能。
  • HAR模型在5 Hz下达到80.2%的准确率,表明低采样率不会显著降低活动识别性能。
  • CNNR方法将所有运动和采样频率下的平均心率估计误差从HeartPy的22.59%降低至20.15%。
  • 经过超参数优化后,CNNR的平均心率估计误差降至13.62%,相比未优化版本降低7个百分点。
  • 优化后的CNNR在5 Hz下仍保持稳健性能,高强度运动下的心率误差为18.90%,步行时为16.47%,与256 Hz结果相当。
  • 在所有测试采样率中,10 Hz采样频率在CNNR模型中实现了最低的心率估计误差,优于15 Hz和30 Hz。

更好的研究,从现在开始

从阅读论文到最终审阅,大幅缩短您的研究时间。

无需绑定信用卡

本解读由 AI 生成,并经人工编辑审核。