Zhan Qi, Wang Wenjin, de Haan Gerard
Department of Electrical and Information Engineering, Hunan University, China.
Remote Sensing Group, Philips Research, The Netherlands.
Biomed Opt Express. 2020 Feb 7;11(3):1268-1283. doi: 10.1364/BOE.382637. eCollection 2020 Mar 1.
Deep learning based on convolutional neural network (CNN) has shown promising results in various vision-based applications, recently also in camera-based vital signs monitoring. The CNN-based photoplethysmography (PPG) extraction has, so far, been focused on performance rather than understanding. In this paper, we try to answer four questions with experiments aiming at improving our understanding of this methodology as it gains popularity. We conclude that the network exploits the blood absorption variation to extract the physiological signals, and that the choice and parameters (phase, spectral content, etc.) of the reference-signal may be more critical than anticipated. The availability of multiple convolutional kernels is necessary for CNN to arrive at a flexible channel combination through the spatial operation, but may not provide the same motion-robustness as a multi-site measurement using knowledge-based PPG extraction. We also find that the PPG-related prior knowledge may still be helpful for the CNN-based PPG extraction, and recommend further investigation of hybrid CNN-based methods that include prior knowledge in their design.
基于卷积神经网络（CNN）的深度学习在各种基于视觉的应用中已显示出有前景的结果，最近在基于摄像头的生命体征监测中也是如此。到目前为止，基于CNN的光电容积脉搏波描记法（PPG）提取一直侧重于性能而非理解。在本文中，我们试图通过实验回答四个问题，旨在随着这种方法越来越受欢迎而增进我们对它的理解。我们得出结论，该网络利用血液吸收变化来提取生理信号，并且参考信号的选择和参数（相位、光谱内容等）可能比预期的更关键。多个卷积核的可用性对于CNN通过空间操作实现灵活的通道组合是必要的，但可能无法提供与使用基于知识的PPG提取的多部位测量相同的运动鲁棒性。我们还发现，与PPG相关的先验知识可能仍然有助于基于CNN的PPG提取，并建议进一步研究在设计中包含先验知识的基于CNN的混合方法。
