Abstract
Purpose :
Microstructural biomarkers in retinal optical coherence tomography (OCT) images are obscured because high-lateral-resolution OCT images are focused to a specific depth and because of ocular aberrations. Adaptive optics overcomes the ocular aberrations, but requires additional expensive hardware. Digital aberration correction (DAC) based on Lissajous-scan OCT and one-dimensional (1D) deconvolution, which is immune to eye motion, has been applied to in vivo human eyes.
Methods :
A healthy eye has been scanned with 1-µm spectral-domain OCT with a probe beam of 3.4-mm diameter and a line rate of 92 kHz. The probe beam scanned with a Lissajous trajectory. Each short segment of the Lissajous pattern has been convolved with 1D complex deconvolution filter. Hence, it improves the lateral resolution mainly along the local scan direction. On the other hand, this method requires the signal phase stability only within the short 1D segment acquired in a few hundred microseconds. The eye motion is corrected by post-processing tailored for the Lissajous scan. During OCT volume reconstruction, multiple 1D deconvolved segments are combined to improve the lateral resolution two-dimensionally.
Results :
Fig. 1 shows the OCT images of an in vivo human eye with and without DAC. En-face images at the inner plexiform layer shows that dot-like patterns become more rounded and sharp. This is also shown in the cross-sectional images. In the nerve fiber layer, the contrast of fiber bundle structure is more apparent.
Conclusions :
The digital aberration correction with Lissajous scan motion correction can restore the lateral resolution of OCT which is deteriorated by defocus and ocular aberrations. This will enable a detection of microstructural biomarkers in OCT without adaptive optics.
This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.