Abstract
Abstract: :
Purpose: The objective of this project to demonstrate the clinical significance of added resolution provided by an adaptive optics retinal imager. Methods:An adaptive optics (AO) fundus imager that is compatible with clinical use has been used to collect images to subjects with diabetes, but no known retinopathy, to determine whether early signs of retinopathy can be detected. In our implementation of an AO fundus imager the low–order correction is achieved through a hardware–based adaptive optic system. Image restoration to remove remaining aberrations is achieved by direct deconvolution using the point spread function (PSF) or a blind deconvolution when the PSF is not available. The PSF is estimated using measured wavefront aberrations. Images taken with the AO fundus imager were used as a basis for a comparison with standard fundus images. Photographs were analyzed and compared by trained ophthalmologists and retinal photograph readers. Results: This system has been tested on six human subjects (three normals and three with diabetes, but no diagnosed retinopathy). Microaneurysms, exudates, and vessel abnormalities can be visualized with the adaptive optics device that cannot be resolved with the standard fundus camera. The individual nerve fiber bundles are easily resolved. Conclusions: This project demonstrated that computer–enhanced, adaptive optic images have greater detail of anatomical and pathological structures.
Keywords: imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • image processing • diabetic retinopathy