December 2013
Volume 13, Issue 15
Free
OSA Fall Vision Meeting Abstract  |   October 2013
A principal component analysis based approach to determine predominant lamina cribrosa beam orientation directly from in vivo images
Author Affiliations
  • Nripun Sredar
    Department of Computer Science, University of Houston, Houston, Texas, USA
  • Kevin M. Ivers
    College of Optometry, University of Houston, Houston, Texas, USA
  • Hope M. Queener
    College of Optometry, University of Houston, Houston, Texas, USA
  • George Zouridakis
    Department of Computer Science, University of Houston, Houston, Texas, USA
    Department of Engineering Technology, University of Houston, Houston, Texas, USA
    Department of Electrical and Computer Engineering, University of Houston, Houston, Texas, USA
  • Jason Porter
    College of Optometry, University of Houston, Houston, Texas, USA
Journal of Vision October 2013, Vol.13, P36. doi:10.1167/13.15.71
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Nripun Sredar, Kevin M. Ivers, Hope M. Queener, George Zouridakis, Jason Porter; A principal component analysis based approach to determine predominant lamina cribrosa beam orientation directly from in vivo images. Journal of Vision 2013;13(15):P36. doi: 10.1167/13.15.71.

      Download citation file:


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

      ×
  • Supplements
Abstract

The lamina cribrosa, a 3D meshwork of collagenous beams within the optic nerve head, is suggested to be the initial site of damage to retinal ganglion cell axons in glaucoma. Previous studies that examined the predominant orientation of laminar beams in histological images have typically been performed on binary segmented raw images using techniques such as mean intercept length (MIL) and radon transform. We have developed a method based on principal component analysis (PCA) that operates directly on grayscale images without the need for binary segmentation. To validate the method, we employed synthetic images of known characteristics to estimate the predominant orientation of overlapping local regions of interest and compared the resulting estimates with the known values for different levels of noise and blur. Overall, the PCA-based method yielded estimates with smaller errors in local orientations (0.47 ± 0.2 degrees) compared to radon transform (14.6 ± 31.5 degrees) and MIL (21.5 ± 45.3 degrees) techniques. The proposed method was subsequently applied to in vivo adaptive optics scanning laser ophthalmoscope images of the anterior laminar surface that were contrast enhanced using a contrast limiting adaptive histogram equalization technique. We found that the local predominant orientation vectors in two normal monkey eyes were mainly radial while the predominant local orientations in the same eyes after induction of experimental glaucoma were circumferential. These findings suggest that the proposed PCA-based technique can be used to accurately estimate local changes in laminar beam orientation in vivo during disease progression.

×
×

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.

×