April 2014
Volume 55, Issue 13
Free
ARVO Annual Meeting Abstract  |   April 2014
Determination of Spectral-Spatial Resolution of Hyperspectral Images For Retinal Imaging Applications
Author Affiliations & Notes
  • Justin S. Baba
    Electrical and Electronics Systems Research Division, Oak Ridge National Laboratory, Oak Ridge, TN
    Institute of Biomedical Engineering (iBME), University of Tennessee, Knoxville, TN
  • Amir H Kashani
    Doheny Eye Institute, University of Southern California, Los Angeles, CA
  • Thomas Paul Karnowski
    Electrical and Electronics Systems Research Division, Oak Ridge National Laboratory, Oak Ridge, TN
  • Gabriel Martin
    Ametek, Reichert Technologies, New York, NY
  • Mark S Humayun
    Doheny Eye Institute, University of Southern California, Los Angeles, CA
  • Footnotes
    Commercial Relationships Justin Baba, None; Amir Kashani, Reichert Technologies (F); Thomas Karnowski, None; Gabriel Martin, Reichert Technologies (E); Mark Humayun, Reichert Technologies (F)
  • Footnotes
    Support None
Investigative Ophthalmology & Visual Science April 2014, Vol.55, 1613. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Justin S. Baba, Amir H Kashani, Thomas Paul Karnowski, Gabriel Martin, Mark S Humayun, imaging/image analysis: non-clinical; Determination of Spectral-Spatial Resolution of Hyperspectral Images For Retinal Imaging Applications. Invest. Ophthalmol. Vis. Sci. 2014;55(13):1613.

      Download citation file:


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

      ×
  • Supplements
Abstract
 
Purpose
 

To determine the spectral-spatial resolution of fundus images obtained with a hyperspectral computed tomographic imaging spectrometer (HCTIS) for retinal imaging applications.

 
Methods
 

HCTIS is a novel methodology that allows simultaneous acquisition of many spectral bands from the retina using standard fundus photography. In order to characterize the spectral-spatial resolution of HCTIS images, Spectralon was used as a reflectance standard and calibrating target. One small rectangular piece of Spectralon (~1mm2) was surgically placed on the rabbit retina in vivo using standard rabbit anesthesia and vitrectomy techniques (Kashani AH et al., PLoS One 2011a,b). Fundus images of the retina spanning 450-660nm were then obtained using a custom HCTIS coupled to a standard fundus camera as previously described (Kashani AH et al., PLoS One 2011a,b). Images were analyzed using Matlab. The spectral-spatial resolution was characterized based on measured distance between the 0.1 and 0.9 values of the normalized line-spread-function (LSF) for normal lines traversing the rectangular sample in both lateral directions. The LSF at each edge, four total, was determined and the mean and standard deviation were computed.

 
Results
 

The overall mean spectral-spatial resolution of the HCTIS using this method is approximately 89.05 um ±1.53 (SD) across the spectral range of 450-660nm. The spectral-spatial resolution of the right, left, top, and bottom edges were 93.28 um ±3.74 (SD), 88.19 um ±3.03 (SD), 82.76 um ±5.34 (SD), and 91.95 um ±3.12 (SD), respectively.

 
Conclusions
 

Fundus images obtained using HCTIS methods can have a spectral-spatial resolution of at least 89um ±1.53 (SD). This is on the scale that is useful for clinical evaluation of retinal lesions. Limitations of this method for establishing spectral-spatial resolution include non-normal illumination and lack of true straightness of the edges of the uniform diffuse reflectance Spectralon sample that was manually cut for surgical placement. Therefore, the results presented represent the higher end of the true spectral-spatial resolution of the HCTIS method.

 
Keywords: 551 imaging/image analysis: non-clinical • 552 imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • 688 retina  
×
×

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.

×