Abstract
Many ocular conditions are linked to changes in the vasculature and circulatory system of the posterior eye. Over the past 25 years, optical coherence tomography (OCT) has revolutionized ocular imaging, providing highly detailed anatomical information on the retina and choroid. However, current clinical OCT devices are unable to quantify blood flow rate or velocity in ocular tissues. This limitation is significant, as such measurements could provide valuable insights for assessing eye health and managing ocular diseases. This research addresses this issue by developing methodologies using a phantom imaging model, feature extraction, and machine learning methods to quantify blood flow rate with clinical OCT imaging. Specifically, it demonstrates that using a range of image features and a linear discriminant model, the flow rate can be classified in OCT images with 90% accuracy by utilizing speckle statistical features. This study represents a novel attempt to automatically quantify flow rate using OCT images and speckle metrics, aiming to enhance the ability of OCT in assessing ocular vasculature.