June 2015
Volume 56, Issue 7
Free
ARVO Annual Meeting Abstract  |   June 2015
Automated ocular surface staining analysis based on a teachable algorithm: smartphone web-based applicayion for dry eye and ocular surface disease
Author Affiliations & Notes
  • Felipe Hering Padovani
    Pontific Catholic University of Campinas, Jundiaí, Brazil
  • Paulo Borges
    Electrical Engineering, Santa Clara University, Santa Clara, CA
  • Kristopher Sanford
    Electrical Engineering, Santa Clara University, Santa Clara, CA
  • Luciana Oharomari
    Ophthalmology, FMRP, Ribeirão Preto, Brazil
  • Eduardo M Rocha
    Ophthalmology, FMRP, Ribeirão Preto, Brazil
  • Paulo Schor
    Ophthalmology, Federal University of São Paulo, São Paulo, Brazil
  • Monica Alves
    Ophthalmolgy, University of Campinas, Campinas, Brazil
  • Footnotes
    Commercial Relationships Felipe Padovani, None; Paulo Borges, None; Kristopher Sanford, None; Luciana Oharomari, None; Eduardo Rocha, None; Paulo Schor, None; Monica Alves, None
  • Footnotes
    Support None
Investigative Ophthalmology & Visual Science June 2015, Vol.56, 4113. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Felipe Hering Padovani, Paulo Borges, Kristopher Sanford, Luciana Oharomari, Eduardo M Rocha, Paulo Schor, Monica Alves; Automated ocular surface staining analysis based on a teachable algorithm: smartphone web-based applicayion for dry eye and ocular surface disease. Invest. Ophthalmol. Vis. Sci. 2015;56(7 ):4113.

      Download citation file:


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

      ×
  • Supplements
Abstract
 
Purpose
 

Dry Eye Disease (DED) and ocular surface disorders diagnosis, scoring and follow up apply vital stainings. However, interpretation is influenced by observer experience and subjectivity. The aim of the present work is to describe a web-based application (app) which runs an algorithm to analyzes images of fluorescein ocular surface staining obtained with the digital camera of a smartphone to quantify, register and provide a follow up tool.

 
Methods
 

Support Vector Machine (SVM) was applied to score the intensity and pattern of fluorescein staining in images of ocular surface obtained from healthy individuals and ocular surface diseases patients using the camera of the smartphone connected to a slit lamp. We are trying an automated approach to reduce the noise and to crop the unwanted background, using a sequence of algorithms from the Open Computer Vision (OpenCV) programming library. The goal is to obtain an image that can be analysed by a standard method which produces a numerical coefficient tha can be compared to clinical results.

 
Results
 

The software determines the optimal hyperplane to provide the most accurate maximum margin to separate the staining area from not one, as seen in punctate keratitis in DED, corneal ulcers and epithelial defects. Gauss filters are being applied to improve the final results. Preliminary results have showed necessity to improve the techniques in order to reduce the noise and crop the image properly and, then, improve the statistical and clinical relevance of the numeric coefficient.

 
Conclusions
 

The present work describes a new tool capable to register and score corneal fluorescein staining. It will be useful for clinical standardization and follow up measurements of DED severity and ocular surface disorders. Regarding practical aspect this tool can run in devices present in the daily clinical practice, the slit lamp and the smartphone, providing a feasible and more precise method to analyse ocular surface staining.  

 
Normal ocular surface: fluorescein staining under cobalt filter, smartphone and slit lamp photograph 10x magnification.
 
Normal ocular surface: fluorescein staining under cobalt filter, smartphone and slit lamp photograph 10x magnification.
 
 
Image editing algorithm prior to staining analysis.
 
Image editing algorithm prior to staining analysis.

 
×
×

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.

×