Abstract
Purpose:
To improve visibility of photoreceptors in capillary shadow region of AOSLO images using characteristics of temporal intensity variation caused by blood cell flow.
Methods:
Our capillary shadow correction method consists of detecting abnormal frames in an AOSLO video, calculating statistical values of temporal intensity variation in each pixel except the abnormal frames, and allocating the pixel value based on the statistical values in each pixel position for producing the resultant contrast enhanced AOSLO image. In this study, we recorded AOSLO videos using Canon prototype AOSLO system with a high wavefront correction efficiency using dual liquid crystal phase modulators. The imaging light wavelength was 840 nm and the frame rate was 32 Hz. The scan area at the parafovea was 1.1 × 1.1° and was sampled at 400 × 400 pixels. The videos were recorded for 2-5 seconds. The proposed adaptive contrast enhancement method was applied to 15 AOSLO videos for 5 healthy subjects including parafoveal capillary regions (leukocyte preferred paths (LPPs) and plasma-gap capillaries (PGCs)) and macular arteriolovenous region. We calculated the spectral power ratio as a function of frequency between the adaptive contrast enhanced images and conventional frame averaged images particularly in LPPs, PGCs and macular arteriolovenous regions.
Results:
The spectral power ratio at a given frequency in LPPs and PGCs of the adaptive contrast enhanced AOSLO images were 1.62 and 1.09 on average respectively, because capillary shadows were corrected by using intensity values when leukocyte or plasma-gap passes. On the other hand, shadows in macular arteriolovenous regions were not corrected because pixel intensity was low in all frames of AOSLO video as the result of erythrocytes next to a leukocyte blocking the imaging light.
Conclusions:
Our method improved contrast of photoreceptors in retinal capillary shadow region of AOSLO images using blood cell flow information.
Keywords: 552 imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) •
688 retina •
549 image processing