June 2017
Volume 58, Issue 8
Open Access
ARVO Annual Meeting Abstract  |   June 2017
The Benefits of Real Time Pupil Tracking on the Quality of the B-Scan
Author Affiliations & Notes
  • Simon Claus Stock
    R&D, Carl Zeiss Meditec, Inc., Dublin, California, United States
    Institute for Information Processing Technologies (ITIV), Karlsruhe Institute of Technology, Karlsruhe, Germany
  • Sophie Kubach
    R&D, Carl Zeiss Meditec, Inc., Dublin, California, United States
  • Patricia Sha
    R&D, Carl Zeiss Meditec, Inc., Dublin, California, United States
  • Jochen Straub
    R&D, Carl Zeiss Meditec, Inc., Dublin, California, United States
  • Wilhelm Stork
    Institute for Information Processing Technologies (ITIV), Karlsruhe Institute of Technology, Karlsruhe, Germany
  • Footnotes
    Commercial Relationships   Simon Stock, Carl Zeiss Meditec, Inc. (C); Sophie Kubach, Carl Zeiss Meditec, Inc. (E); Patricia Sha, Carl Zeiss Meditec, Inc. (C); Jochen Straub, Carl Zeiss Meditec, Inc. (E); Wilhelm Stork, None
  • Footnotes
    Support  None
Investigative Ophthalmology & Visual Science June 2017, Vol.58, 660. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to Subscribers Only
      Sign In or Create an Account ×
    • Get Citation

      Simon Claus Stock, Sophie Kubach, Patricia Sha, Jochen Straub, Wilhelm Stork; The Benefits of Real Time Pupil Tracking on the Quality of the B-Scan. Invest. Ophthalmol. Vis. Sci. 2017;58(8):660.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose : In retinal imaging eye instruments, such as Optical Coherence Tomography (OCT) imaging, the pupil entry
point is crucial for the quality of the scan. This is of particular importance for subjects with small pupils (less than 2 mm) or those with cataracts. The purpose of this study is to demonstrate the accuracy of a pupil tracking algorithm that could be used to maintain a consistent pupil entry point in an OCT instrument.

Methods : A real-time pupil tracking algorithm was tested offline on a set of iris images previously acquired on a CIRRUS™ HD-OCT 5000 (ZEISS, Dublin, CA) instrument. The data set consists of n=295 images with various levels of image quality. In order to rate and validate the pupil tracking algorithm, we first defined the ground truth manually for the pupil centers and boundaries (see Figure 1). The pupil tracking algorithm follows a gradient approach and adds further filtering to better exclude artifacts such as those caused by the eye lid and reflections. The output of the algorithm is a least-squares and RANSAC fitted ellipse (see Figure 1).

Results : The pupil detection algorithm is able to determine the center of the pupil with a mean accuracy of 7.65 pixels
or 230 μm (calculated over the 295 individual test cases), where the accuracy is defined by the ground truth.
The study was performed on 3 subjects known to have a small pupil which makes the quality of the B-Scan more
sensitive to misalignments. In Figure 2 it can be seen that a misaligned B-Scan (B) has a reduced signal to noise ratio in comparison with the optimal alignment (A).

Conclusions : We demonstrated the benefit pupil detection has on the quality of the B-Scan. The proposed method has the potential to be implemented along with a closed loop control system for pupil alignment and tracking. A workflow based on pupil tracking may improve the measurement repeatability on OCT instruments and reduce the interaction between the operator and the instrument.

This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.

 

Figure 1: A: Original Iris Image. B: Ground Truth. C: Ground truth for pupil boundaries and pupil center. D: Output of the algorithm: Feature points on pupil. E: Fitted ellipse. F: Comparison of ground truth and algorithm.

Figure 1: A: Original Iris Image. B: Ground Truth. C: Ground truth for pupil boundaries and pupil center. D: Output of the algorithm: Feature points on pupil. E: Fitted ellipse. F: Comparison of ground truth and algorithm.

 

Figure 2: Difference in B-Scan quality between a correctly centered and misaligned pupil. The misaligned pupil causes a reduction of the signal to noise ratio as it can be seen on the left part of the B-Scan.

Figure 2: Difference in B-Scan quality between a correctly centered and misaligned pupil. The misaligned pupil causes a reduction of the signal to noise ratio as it can be seen on the left part of the B-Scan.

×
×

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.

×