Abstract
Purpose :
To demonstrate that a hybrid optoelectronic device that combines imaging Scanning Laser Ophthalmoscope (SLO) with ultrafast SLO based retinal tracker allows for generation of high resolution SLO images with reduced imaging artifacts in patients with fixation problems.
Methods :
The custom built system optically combines imaging SLO with ultrafast and low field of view SLO device that provides over 1200 retinal images per second allowing detection of ocular motion components, namely drift, tremor, microsaccades and saccades in the living human eye. The latter SLO uses two dimensional microelectromechanical systems (MEMS) scanner oscillating at ~20 kHz (fast axis) and ~600 Hz (slow axis) to acquire image (256 x 16 pixels each) of 3°x3° region of the retina. Imaging SLO device operates at 30 frames to provide images (800x600 pixel each) of ~24°x15° region of the retina. All the images are captured by a custom-built electronics, transferred to synchronizing computer, the displacements between consecutive retinal images caused by the eye motion is found using specifically designed algorithms based on image correlation and used to reconstruct a single high resolution and highly sampled image from a number of regular images from slow SLO.
Results :
The ability to provide high resolution SLO images using retinal trajectory reconstructed from ultrafast retinal tracker is proven on 10 eyes of 6 healthy volunteers and 2 patients with diabetic retinopathy aged between 28 and 64 y.o. The trajectory estimation precision is less than 4 mm. The residual image displacements, related to image distortions (caused by ocular aberrations, scanning arrangement of the imaging setup etc.) are corrected numerically. To induce retinal motion the fixation target was dynamically displaced forcing patients to change gaze direction during data acquisition.
Conclusions :
The results show that SLO based ultrafast imaging is able to provide retinal images allowing for precise retinal tracking, that can be used for enhancing imaging field of view or to reconstruct the trajectory of eye movements occurring during retinal imaging. The first allows for reconstruction of high resolution images with increased signal to noise ratio and reduced imaging artifacts even in patients with fixation disorders. The latter can be additionally used for early detection of eye diseases related with eye movements disorders.
This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.