July 2019
Volume 60, Issue 9
Open Access
ARVO Annual Meeting Abstract  |   July 2019
Aberration estimation by computational adaptive optics to improve the quality of images containing structured illumination
Author Affiliations & Notes
  • Jyoti Paul
    School of Engineering and IT, The University of New South Wales, Canberra, Australian Capital Territory, Australia
  • Andrew J Lambert
    School of Engineering and IT, The University of New South Wales, Canberra, Australian Capital Territory, Australia
  • Footnotes
    Commercial Relationships   Jyoti Paul, None; Andrew J Lambert, None
  • Footnotes
    Support  None
Investigative Ophthalmology & Visual Science July 2019, Vol.60, 607. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Jyoti Paul, Andrew J Lambert; Aberration estimation by computational adaptive optics to improve the quality of images containing structured illumination. Invest. Ophthalmol. Vis. Sci. 2019;60(9):607.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose : Structured fringe pattern projected on human retina is a component for determining super-resolved imagery. However, identifying fringes in an image, especially with higher frequency is difficult due to the eye’s optical aberrations. We demonstrate a computational adaptive optics (CAO) technique to obtain and remove aberrations from imaging of retina based on a method developed by Hillmann et al, Scientific Reports, 6, 2016, for OCT images, which provides an improved resolution even from a single image, acquired with the structured illumination.

Methods : Interference fringes are directly produced at the central retina with a modified Lotmar interferometer without dilating the pupil. The aberrated phase is then iteratively determined and described by Zernike polynomials. A CAO correction technique is applied which uses an inverse optimization method involving the fast Fourier transformation of the estimated phase. The improvement of image quality is evaluated by normalized image intensity and observing the Shannon entropy to converge this algorithm. Finally, the image is reconstructed according to the optimized information.

Results : Using the imagery of Zhu et al, Journal of Vision, 16(10), 2016, four visual acuity levels (0.05, 0.10, .015, and 0.20) are registered for each fringe orientation (45°, 90°, 135°, and 180°). The laser power on the cornea is 28.5µW with a maximum output of 8.03mW. Aberrations up to 8th radial degree are considered to correct the phase error of fringes using the Zernike polynomials and the best result is obtained for the 7th radial degree. The information content both in bright and dark regions are estimated through Shannon entropy that determines minimum aberration. Thus, the illumination fringes are more easily identified leading to optimized super-resolution image reconstruction.

Conclusions : The fringes from the illumination are hard to determine because of the aberrations. We evaluated whether images with these fringes, which would have spatially varying entropy because of the dark parts in the fringe spacing, could still inform the algorithm determining aberrations. This approach can estimate and correct the phase aberrations induced with the structured illumination both for bright and dark regions even with only one depth image plane, therefore, can be used to generate super-resolution retinal images.

This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.

×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×