May 2006
Volume 47, Issue 13
Free
ARVO Annual Meeting Abstract  |   May 2006
Automatic Tracing of Subbasal Nerves in Confocal Microscopy Images
Author Affiliations & Notes
  • A. Ruggeri
    Dept of Information Engineering, University of Padova, Padova, Italy
  • F. Scarpa
    Dept of Information Engineering, University of Padova, Padova, Italy
  • E. Grisan
    Dept of Information Engineering, University of Padova, Padova, Italy
  • Footnotes
    Commercial Relationships  A. Ruggeri, Nidek Technologies, C; F. Scarpa, None; E. Grisan, None.
  • Footnotes
    Support  None
Investigative Ophthalmology & Visual Science May 2006, Vol.47, 1338. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      A. Ruggeri, F. Scarpa, E. Grisan; Automatic Tracing of Subbasal Nerves in Confocal Microscopy Images . Invest. Ophthalmol. Vis. Sci. 2006;47(13):1338.

      Download citation file:


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

      ×
  • Supplements
Abstract
 
Purpose:
 

We propose a new, fully automatic procedure for the recognition and tracing of subbasal nerves in confocal microscopy images of the cornea.

 
Methods:
 

The ConfoScan4 corneal confocal microscope (Nidek Technologies, Italy) allows the rapid and non–invasive acquisition also of subbasal layer images. These images are at first processed to normalize image luminosity and contrast. Seed points are then detected all over the image, to be used as starting points for the nerve tracing algorithm. Nerve segments are recognized on the entire image by a tracking–type algorithm and techniques specific for this kind of images have been implemented to increase the algorithm sensitivity. An ad–hoc post–processing module was designed to remove false recognitions, mainly due to keratocytes, by recognizing very bright blobs in the image and selectively removing recognitions inside these blobs. Finally, the procedure links sparse segments into as much as possible continuous nerve structures. Significant quantitative clinical parameters can then be easily derived, such as total length of nerves in the image, or nerve density (length of nerve per mm2), or nerve tortuosity. A prototype of the algorithm was implemented in the Matlab® language and run on a personal computer.

 
Results:
 

A preliminary evaluation was performed on a data set containing 40 images acquired in normal subjects. The percent of traced nerves (with respect to manual tracing) was on average 90%; the rate of false positive recognitions was on average 8%. A representative example of the results obtained is shown in the enclosed figure. Running times of the prototype were 4–5 minutes per image, but its final implementation with a more efficient computer language, e.g. C++, will allow reducing running times to tens of seconds.

 
Conclusions:
 

Issues still to be considered are the balance between sensitivity and specificity of the recognitions from a clinical point of view and the acceptability of allowing some manual editing to improve them.  

 
Keywords: imaging/image analysis: clinical • cornea: epithelium • microscopy: confocal/tunneling 
×
×

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.

×