Speckle pattern analysis has been found by many researchers to be applicable to remote sensing of various biomedical parameters. This paper shows how analysis of dynamic differential speckle patterns scattered from subjects’ sclera illuminated by a laser beam allows extraction of micro-saccades movement in the human eye. Analysis of micro-saccades movement using advanced machine learning techniques based on convolutional neural networks offers a novel approach for non-contact assessment of human blood oxygen saturation level (SpO2). Early stages of hypoxia can rapidly progress into pneumonia and death, and lives can be saved by advance remote detection of reduced blood oxygen saturation.

中文翻译：

---

通过跟踪微扫视模式中的异常，对血氧饱和度进行远程光子传感

许多研究人员发现散斑图样分析适用于各种生物医学参数的遥感。本文显示了如何分析由激光束照射的受试者巩膜散布的动态差异斑点图案，从而可以提取人眼中的微扫视运动。使用基于卷积神经网络的先进机器学习技术对微扫视运动进行分析，为人类血液氧饱和度水平（SpO2）的非接触式评估提供了一种新颖的方法。缺氧的早期阶段可以迅速发展为肺炎和死亡，并且可以通过预先远程检测降低的血氧饱和度来挽救生命。