September 2016
Volume 57, Issue 12
Open Access
ARVO Annual Meeting Abstract  |   September 2016
Automated quantitation of the endothelium alterations in Eye Bank processed corneas.
Author Affiliations & Notes
  • Brian Madow
    Ophthalmology, University of South Florida, Tampa, Florida, United States
  • Nicholas Sprehe
    Lions Eye Institute, Tampa, Florida, United States
  • Sung Lee
    Minnesota Lions Eye Bank, St Paul, Minnesota, United States
  • Patrick Gore
    Lions Eye Institute, Tampa, Florida, United States
  • Stephen Kaufman
    Minnesota Lions Eye Bank, St Paul, Minnesota, United States
  • Jackie Malling
    Minnesota Lions Eye Bank, St Paul, Minnesota, United States
  • Veronique Grimes
    Minnesota Lions Eye Bank, St Paul, Minnesota, United States
  • Mitchell D McCartney
    Lions Eye Institute, Tampa, Florida, United States
  • Footnotes
    Commercial Relationships   Brian Madow, None; Nicholas Sprehe, None; Sung Lee, None; Patrick Gore, None; Stephen Kaufman, None; Jackie Malling, None; Veronique Grimes, None; Mitchell McCartney, None
  • Footnotes
    Support  None
Investigative Ophthalmology & Visual Science September 2016, Vol.57, 1931. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Brian Madow, Nicholas Sprehe, Sung Lee, Patrick Gore, Stephen Kaufman, Jackie Malling, Veronique Grimes, Mitchell D McCartney; Automated quantitation of the endothelium alterations in Eye Bank processed corneas.. Invest. Ophthalmol. Vis. Sci. 2016;57(12):1931.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose : To design and validate image processing algorithm for automated evaluation of corneal tissue processed at the Eye Bank. The quality control of the corneal tissue supplied to the surgeons for transplantation is of paramount importance. Currently the existing methods are subjective and time consuming. We sought to develop automated image processing application for rapid and objective evaluation of the preprocessed donor cornea disks.

Methods : Processed donor cornea disks were stained using a modification of Park et al. (2012). High magnification color photographs were obtained. Images were stored on disk for further evaluation. Image processing algorithms were developed using MathLab software. Each image was preprocessed for brightness and contrast and the image size was equalized for all images. Color channels were extracted for the three main colors and analyzed. Each channel was post processed separately. Multiple runs were performed on several pilot images to determine the desired threshold values. Each image analyzed by the algorithm was evaluated for adequate feature extraction as compared to the original color photograph. The second part of the algorithm aimed to extract the size of the corneal disk by using image segmentation. The size of the disk was expressed in number of pixels. Finally the software automatically calculated the size of the extracted corneal features as a percentage of each cornea.

Results : 34 images processed and received from 2 separate Eye Bank facilities were used to design and validate the algorithm. Satisfactory automatic feature extraction of the corneal staining was achieved 100 %. In 5 cases additional image noise due to under- or overexposure was manually filtered outside the area of feature extraction. The algorithm was able to segment the cornea dimensions in 100% of the images. Repeating the process for all images yielded 100% reliability. The total size of the corneal features attributed to corneal alteration varied from 8.53 - 45.83 % with median of 22.46.

Conclusions : Software image algorithm for automated quantitation of the endothelium viability from Eye Bank processed corneas was successfully developed and validated. It yielded 100% extraction rate and reliability. It was found that the noise contained in the images is a function of the photographic process which needs further optimization.

This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.

×
×

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.

×