April 2014
Volume 55, Issue 13
ARVO Annual Meeting Abstract  |   April 2014
Registration accuracy of NIDEK MP-1 micro-perimeter in ‘follow-up’ testing
Author Affiliations & Notes
  • Arunkumar Krishnan
    Optometry, Univ of Houston College of Optometry, Houston, TX
  • Nimesh Bhikhu Patel
    Optometry, Univ of Houston College of Optometry, Houston, TX
  • Scott B Stevenson
    Optometry, Univ of Houston College of Optometry, Houston, TX
  • Harold E Bedell
    Optometry, Univ of Houston College of Optometry, Houston, TX
  • Footnotes
    Commercial Relationships Arunkumar Krishnan, None; Nimesh Patel, None; Scott Stevenson, None; Harold Bedell, None
  • Footnotes
    Support None
Investigative Ophthalmology & Visual Science April 2014, Vol.55, 6113. doi:
  • Views
  • Share
  • Tools
    • Alerts
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Arunkumar Krishnan, Nimesh Bhikhu Patel, Scott B Stevenson, Harold E Bedell, None; Registration accuracy of NIDEK MP-1 micro-perimeter in ‘follow-up’ testing. Invest. Ophthalmol. Vis. Sci. 2014;55(13):6113.

      Download citation file:

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

  • Supplements

Purpose: The follow-up test option of micro-perimeters like the MP-1 provides functional information about central retinal disease progression, by tracking sensitivity changes over time at a specific set of locations. Here, we investigate the accuracy of MP-1 in testing the same retinal locations during baseline and follow-up testing.

Methods: Data were collected from one eye of 11 normal subjects (5 older, >50 years and 6 younger) after pupillary dilation. A custom array of 36 Goldman-size-II (diameter = 13 arc min) test points in a 1.2° square grid was used to screen the retinal sensitivity at a non-foveal location while subjects fixated at the center of a 10-deg radius circle. Both baseline and follow-up testing were done on the same day, using 3 different IR illuminance levels. Two regions of interest (ROIs) with maximum ROI indices were selected in the baseline image, and the MP-1 used them to register images and position the test array for follow-up testing. IR image pairs (768x576 pixels) returned by MP-1 were registered using a Generalized Dual Bootstrap - Iterative Closest Point (GDB-ICP) algorithm (Yang et al, 2007). The algorithm matches key extracted features by generating a similarity transformation first locally within a small region, which then expands slowly to cover the whole image. Affine transformation was applied and the coefficients returned by the algorithm were used to transform the follow-up test points. Mean absolute offsets from GDP-ICP-defined registration in the horizontal and vertical directions were computed as the difference between corresponding test point locations in each pair of baseline and follow-up images.

Results: The average (N=11, ±SD) absolute horizontal offsets were 4.4±3.3, 2.5±2.6 and 4.2±4.2 arc min, respectively for the low, medium and high IR settings. The corresponding values for vertical offsets were 3.7±2.4, 3.3±2.3 and 3.9±2.9 arc min. The average offsets across illuminance levels were not significantly different for the younger and older subjects.

Conclusions: The NIDEK MP-1 registers the IR images during follow-up testing with an accuracy close to one pixel (1 pixel = ~4 arc min). Registration is equally good for both younger and older subjects and does not differ significantly across IR illuminance settings. The MP-1 can be used to assess essentially identical retinal locations repeatedly with a dense sampling grid.

Keywords: 642 perimetry • 550 imaging/image analysis: clinical • 552 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.