Visualization and Analysis of Perfusion Cluster Velocity From a Large Area Near-Infrared Image Sensor

Abstract

Flexible, large area, image sensors (FLA-IS) can enable the capture of additional biomarkers not found in current wearable devices with discrete optical sensors. To demonstrate, this study utilizes an in-house, near-infrared sensitive FLA-IS system to collect data from multiple volunteers on a finger test site. A custom processing algorithm is applied to reveal subcutaneous perfusion movement and quantify that into a blood flow velocity equivalent biomarker: perfusion cluster velocity (PcV). The letter analyzes the variability and repeatability of all the processed signals by tracking features representing duration between selected points of interest for PcV and reference photoplethysmography (PPG) waveforms per cardiac cycle. The resulting analyses show 24% improvement in dynamic range and 20% less cycle-to-cycle variation for PcV, and PcV, PPG features compared to PPG-only features, demonstrating FLA-IS's potential for advanced biomarker measurements.