May 2005
Volume 46, Issue 13
ARVO Annual Meeting Abstract  |   May 2005
Image Deconvolution Improves the Repeatability of Topographic Height Measurements in Scanning Laser Tomography
Author Affiliations & Notes
  • A.J. Patterson
    School of Science, Nottingham Trent University, Nottingham, United Kingdom
  • D.P. Crabb
    School of Science, Nottingham Trent University, Nottingham, United Kingdom
  • D.F. Garway–Heath
    Glaucoma Research Unit, Moorfields Eye Hospital, London, United Kingdom
  • Footnotes
    Commercial Relationships  A.J. Patterson, None; D.P. Crabb, None; D.F. Garway–Heath, None.
  • Footnotes
    Support  None.
Investigative Ophthalmology & Visual Science May 2005, Vol.46, 2508. doi:
  • Views
  • Share
  • Tools
    • Alerts
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      A.J. Patterson, D.P. Crabb, D.F. Garway–Heath; Image Deconvolution Improves the Repeatability of Topographic Height Measurements in Scanning Laser Tomography . Invest. Ophthalmol. Vis. Sci. 2005;46(13):2508.

      Download citation file:

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

  • Supplements

Abstract: : Purpose:To evaluate Maximum Likelihood (ML) Blind Deconvolution as a technique for improving the repeatability of topographic height measurements obtained from scanning laser tomography [Heidelberg Retinal Tomograph (HRT)] Methods: ML Blind Deconvolution is an image processing technique which estimates the original scene from a degraded image using a maximum likelihood probability. This technique has been used in confocal scanning laser microscopy to remove out of focus haze and axial smearing in 3D confocal image stacks. ML Blind Deconvolution requires no prior estimation of the Point Spread Function (PSF) as opposed to other classical linear deconvolution methods. Instead, based on the optical setup of the confocal scanning device, image digitalization spacing and noise levels the algorithm estimates an initial PSF, and iteratively proceeds to a solution. The approach was evaluated in a test–retest series of HRT images from 40 ocular hypertensive and glaucomatous patients with varying degrees of media opacity. Three mean topography images were calculated from nine confocal image stacks (three for each mean) using HRT software (v2.01b). Repeatability of topographic height measurements was quantified by calculating a mean pixel height standard deviation (MPHSD) between the three mean topography images. This whole process was repeated with the confocal stacks deconvolved using AutoDeblur® v7.3 [AutoQuant, NY, USA] software. Results: The deconvolved topography images yielded an improvement in repeatability of topographic height measurements in 33 out of 40 subjects. The mean improvement in MPHSD was 1.8µm (P=0.028). In 10 subjects with an initial MPHSD above 30µm the mean improvement was 6.3µm (P=0.004). There is a strong association between the improvement of MPHSD and the average level of the MPHSD (P=0.012.) Conclusions: Our study demonstrates the utility of the ML Blind Deconvolution algorithm in improving HRT topographic image quality. This improvement is particularly notable in subjects with media opacity and poor image quality. Further work will establish if this technique improves the repeatability of neuro–retinal rim measurements which typically have low signal strength in Scanning Laser Tomography.

Keywords: image processing • optic disc • imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) 

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.