May 2004
Volume 45, Issue 13
ARVO Annual Meeting Abstract  |   May 2004
Automatic biometrics of the anterior eye
Author Affiliations & Notes
  • D.R. Iskander
    School of Optometry, Queensland University of Technology, Brisbane, Australia
  • M.J. Collins
    School of Optometry, Queensland University of Technology, Brisbane, Australia
  • B. Davis
    School of Optometry, Queensland University of Technology, Brisbane, Australia
  • Footnotes
    Commercial Relationships  D.R. Iskander, None; M.J. Collins, None; B. Davis, None.
  • Footnotes
    Support  none
Investigative Ophthalmology & Visual Science May 2004, Vol.45, 2779. doi:
  • Views
  • Share
  • Tools
    • Alerts
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      D.R. Iskander, M.J. Collins, B. Davis; Automatic biometrics of the anterior eye . Invest. Ophthalmol. Vis. Sci. 2004;45(13):2779.

      Download citation file:

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

  • Supplements

Abstract: : Purpose:(i) Develop a computer–based procedure for accurate determination of two–dimensional characteristics of the anterior eye including the pupil, limbus, and lid aperture from digital images. (ii) Illustrate the potential application of the procedure for high–speed filming, wavefront sensing, and in a study of the palpebral aperture in Caucasians and Asians. Methods: A specialized algorithm was developed. First, a procedure determines the areas of the image where the transition from the iris to the sclera is clearly visible. The limbus outline is then determined by iterative gradient techniques in polar coordinates. Similar techniques are used to establish the pupil outline and its relative position to the limbus. To evaluate the proposed algorithm, images from a standard digital camera, a high–speed camera, and the pupil viewer of a wavefront sensor were used. Many of the images were lacking contrast due to the low level of light during the acquisition. For the purpose of measuring the palpebral aperture, additional lines are drawn from the centers of pupil and limbus to assist the operator in determining the eyelid positions. Images of 25 young Caucasian and 24 Asian eyes were acquired with a digital camera in primary gaze. For each of the images the vertical palpebral aperture was estimated along with the pupil and limbus sizes. Results: (High–speed Filming) We used a high–speed camera with 250 Hz sampling rate. The proposed procedure performed well. However, when the eyelids obscure the pupil during blinking, the algorithm slightly overestimates or underestimates the true pupil size. (Wavefront Sensing) Since the measurement of the wavefront aberration is often performed in scotopic conditions, it is of interest to relate the wavefront measurement back to that of photopic pupil size. However, the pupil size and center changes with illumination so these two parameters are measured relative to the limbus outline. The image from the pupil viewer of the wavefront sensor can be used for this task. Despite the very low intensity of the image, the algorithm can successfully detect the limbus and pupil outlines. (Palpebral Aperture) The average limbus diameter was found to be 12.16±0.36 for Caucasians and 11.7±0.50mm for Asians. In the primary gaze, the average vertical palpebral aperture was 10.99±1.33mm and 8.91±1.72mm, respectively. Conclusions: The technique was able to automatically relate the pupil position to the limbus, and to accurately locate the relative lid positions. In the illustrative study, we found that the limbus diameter as well as the vertical palpebral aperture of the Caucasian subjects was significantly larger than that of Asian subjects.

Keywords: image processing • imaging/image analysis: clinical • pupil 

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.