文章摘要

文章信息/Info

Title:: A touchless SpO2 estimation method based on the analysis of dorsal hand video streams

作者:: 白培瑞1; 轩辕梦玉1; 傅颖霞2; 袁梦1; 刘国忠3; 刘庆一1△; （1.山东科技大学电子信息工程学院，青岛 266590；2.山东科技大学能源与矿业工程学院，青岛 266590；3.山东科技大学体育学院，青岛 266590）

Author(s):: BAI Peirui1; XUANYUAN Mengyu1; FU Yingxia2; YUAN Meng1; LIU Guozhong3; LIU Qingyi1; （1.College of Electronic and Information Engineering, Shandong University of Science and Technology, Qingdao 266590, China；2.College of Energy and Mining Engineering, Shandong University of Science and Technology, Qingdao 266590；3.College of Physical Education, Shandong University of Science and Technology, Qingdao, 266590）

Keywords:: Peripheral oxygen saturation; Dorsal hand video analysis; Eulerian video magnification; Smart health monitoring

摘要:: 为提高外周血氧饱和度（SpO2）的测量效率与准确率，本研究提出了一种基于手背视频的非接触式SpO2测量方法。通过欧拉视频放大中的高斯金字塔与放大因子的组合改进，减少光线噪声的干扰，有效放大手背视频中的感兴趣区域。然后，提取放大后视频帧的红、蓝通道，并采用吸光度比值法计算SpO2值。最后，在自制数据集上进行了两种呼吸状态下的实验验证。实验结果表明，所提方法SpO2测量的平均均方根误差（RMSE）可达0.48%，与指夹式脉搏血氧仪相比，表现出较高的一致性，且发现性别对SpO2的测量误差有一定影响。该方法可便捷地移植到智能手机等简易视频设备，对SpO2的日常监测具有良好的应用前景。

Abstract:: To improve the measurement efficiency and accuracy of peripheral blood oxygen saturation (SpO2), a touchless peripheral SpO2 estimation method was proposed based on the analysis of dorsal hand videos. By improving the Gaussian pyramid and magnification factor in Eulerian video magnification, the interference of light noise was reduced, and the region of interest (ROI) of the frames in the dorsal hand video was magnified effectively. Then, the red and blue channels of the magnified video frames were extracted and the absorbance ratio method was used to calculate SpO2 values. At last, experimental validation was performed on self-produced dataset under two respiratory states. The experimental results showed that the proposed method achieved an average root mean square error (RMSE) of 0.48% for SpO2 measurements, which exhibited a high degree of consistency when compared to finger-clip pulse oximetry, and gender was found to have effects on the measurement error of SpO2. It is convenient to embed the proposed method into portable video capturing equipment such as smartphone, and has a good application prospect for the daily monitoring of SpO2.