June 2013
Volume 54, Issue 15
Free
ARVO Annual Meeting Abstract  |   June 2013
Evaluating Lanczos2 Image Filtering for Visual Acuity In Simulated Prosthetic Vision
Author Affiliations & Notes
  • Paulette Lieby
    Computer Vision Research Group, NICTA, Canberra, ACT, Australia
    College of Engineering and Computer Science, Australian National University, Canberra, ACT, Australia
  • Adele Scott
    Computer Vision Research Group, NICTA, Canberra, ACT, Australia
    College of Engineering and Computer Science, Australian National University, Canberra, ACT, Australia
  • Nick Barnes
    Computer Vision Research Group, NICTA, Canberra, ACT, Australia
    College of Engineering and Computer Science, Australian National University, Canberra, ACT, Australia
  • Ashley Stacey
    Computer Vision Research Group, NICTA, Canberra, ACT, Australia
    College of Engineering and Computer Science, Australian National University, Canberra, ACT, Australia
  • Lauren Ayton
    Centre for Eye Research, Melbourne, VIC, Australia
  • Janine Walker
    Computer Vision Research Group, NICTA, Canberra, ACT, Australia
    Centre for Mental Health Research, Australian National University, Canberra, ACT, Australia
  • Footnotes
    Commercial Relationships Paulette Lieby, None; Adele Scott, None; Nick Barnes, None; Ashley Stacey, None; Lauren Ayton, None; Janine Walker, None
  • Footnotes
    Support None
Investigative Ophthalmology & Visual Science June 2013, Vol.54, 1049. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Paulette Lieby, Adele Scott, Nick Barnes, Ashley Stacey, Lauren Ayton, Janine Walker; Evaluating Lanczos2 Image Filtering for Visual Acuity In Simulated Prosthetic Vision. Invest. Ophthalmol. Vis. Sci. 2013;54(15):1049.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose: Optimising visual acuity is an important challenge in prosthetic vision. With a small pupil size, the eye optics can result in low-pass filtering of the retinal image at the Nyquist frequency with respect to the photoreceptor density so that aliasing is not apparent. The Lanczos2 filter offers a better compromise than other image filters in reducing aliasing while retaining image sharpness. We evaluated the performance of Lanczos2 image filtering for visual acuity in simulated prosthetic vision using a validated clinical measure.

Methods: 26 normally-sighted (20/20) adult participants each completed on average 22 trials of the Freiburg Visual Acuity Test (FrACT) during one experimental session. The test material consisted of a black background with a white Landolt C whose gap orientation was presented randomly in any of eight directions. The participants had to determine the gap orientation using a keypad. Visual acuity was automatically computed by the FrACT program. Before presentation on a computer monitor, the test material was processed to simulate prosthetic vision using 98 phosphenes rendered over eight levels of brightness. Interphosphene spacing was 1.79 of visual field. A Lanczos2 filter was applied to the test material image prior to sampling. Five cut-off frequencies, fc=a*fN, a={2.0, 1.33, 1.0, 0.80, 0.66}, fN= Nyquist frequency, and a no-filtering case were used. Participants were masked to the presentation order of the six filtering conditions which were randomly allocated and equally represented in the session. Institutional Ethics Board approval was obtained.

Results: The highest mean VA scores were obtained for the cut-off frequencies of 1.33*fN (n=96, mean logMAR 1.655) and fN (n=96, mean logMAR 1.644) which were not significantly different (p=.985). However both were significantly higher than the mean VA scores obtained for the three other cut-off frequencies and no-filtering (p<.001). The lowest mean VA score was obtained for no-filtering (n=97, mean logMAR 1.816). All the mean VA scores obtained were significantly higher than the highest score predicted by the Nyquist theorem (i.e., 2.031 logMAR) at this phosphene spacing.

Conclusions: Lanczos2 filtering with cut-off at or just above the Nyquist frequency may be used as an effective approach for enhancing visual acuity under conditions of limited resolution and dynamic range characterising current prosthetic vision devices.

Keywords: 549 image processing • 754 visual acuity  
×
×

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.

×