June 2017
Volume 58, Issue 8
ARVO Annual Meeting Abstract  |   June 2017
Assessment of second order statistics for evaluating the spatial distribution of the cones in adaptive optics images of the human retina
Author Affiliations & Notes
  • Daniela Giannini
    IRCCS Fondazione GB BIETTI, Rome, Italy
  • Marco Lombardo
    IRCCS Fondazione GB BIETTI, Rome, Italy
  • sebastiano serrao
    IRCCS Fondazione GB BIETTI, Rome, Italy
  • Giuseppe Lombardo
    Istituto per i Processi Chimico-Fisici, Consiglio Nazionale delle Ricerche, Messina, Italy
    Vision Engineering Italy srl, Rome, Italy
  • Footnotes
    Commercial Relationships   Daniela Giannini, None; Marco Lombardo, None; sebastiano serrao, None; Giuseppe Lombardo, None
  • Footnotes
    Support  None
Investigative Ophthalmology & Visual Science June 2017, Vol.58, 301. doi:
  • Views
  • Share
  • Tools
    • Alerts
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Daniela Giannini, Marco Lombardo, sebastiano serrao, Giuseppe Lombardo; Assessment of second order statistics for evaluating the spatial distribution of the cones in adaptive optics images of the human retina. Invest. Ophthalmol. Vis. Sci. 2017;58(8):301.

      Download citation file:

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

  • Supplements

Purpose : To evaluate the distribution of the cones in adaptive optics (AO) images of the retinal photoreceptor mosaic with second order spatial statistics functions.

Methods : The second order statistics included the G(r), K(r), and g(r) functions. The G(r) function is the cumulative frequency distribution of the nearest neighbor distances, the K(r) function is the cumulative version of the density recovery profile and g(r) is the pair correlation function, which measures the probability of finding a cone at distance of r from a reference cone. Each function was validated using a simulated mosaic, which was generated with a random model (Inhibition Poisson point process) having a-priori knowledge of minimal distance between the cones at the given retinal eccentricity. In addition, AO images of the parafoveal cone mosaic were acquired in twenty healthy subjects and ten patients with retinal disorders. All the metrics were calculated using 204x204 µm sampling area. The random cone mosaic was used as a reference system to characterize spatial correlations in the distributions of the cones in real retinal mosaics. The metrics used included the sum of the Euclidean differences between the curve profile of each function. In addition, the G(r) function was modeled as a stretch logistic function and the slope was used for comparison between healthy and diseased retina.

Results : The departures from simulated cone mosaics of the K(r) and g(r) functions showed significant differences between healthy eyes and retinal diseases (P=0.007 and P=0.002, respectively). The slope of the modeled G function showed significant differences between healthy and diseased retina (P=0.02).

Conclusions : Deviations of spatial statistics of the cone coordinates in real retinal mosaics from those of the random model based on inhibition Poisson point process provide a measure of the degree of spatial correlation between cones. The G(r) and g(r) functions show high sensitivity to discriminate between healthy and diseased retinas. The K(r) function provides a measure to evaluate the cluster pattern in the cone mosaic.

This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.


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.