Automated image assessment of posterior capsule opacification using Hölder exponents

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

2 Scopus citations

Abstract

Posterior Capsule Opacification (PCO) remains to be the most common complication of cataract surgery after intraocular lens implantation. Though several strategies have been suggested for the prevention of PCO, a standard PCO quantification system is required to reliably assess the effectiveness of these strategies. This paper proposes a method based on computation of Hölder exponents to quantify the amount of PCO in the digital image. PCO areas are effectively detected and classified according to their severity using histogram-based thresholding on Hölder exponent image. This method is implemented in Matlab and verified on real PCO images. The results show a high correlation of 83% between the computed PCO scores and the clinical grades, as well as demonstrate the robustness of the proposed system to monotonic illumination variations.

Original languageBritish English
Title of host publication2013 IEEE 20th International Conference on Electronics, Circuits, and Systems, ICECS 2013
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages538-541
Number of pages4
ISBN (Print)9781479924523
DOIs
StatePublished - 2013
Event2013 IEEE 20th International Conference on Electronics, Circuits, and Systems, ICECS 2013 - Abu Dhabi, United Arab Emirates
Duration: 8 Dec 201311 Dec 2013

Publication series

NameProceedings of the IEEE International Conference on Electronics, Circuits, and Systems

Conference

Conference2013 IEEE 20th International Conference on Electronics, Circuits, and Systems, ICECS 2013
Country/TerritoryUnited Arab Emirates
CityAbu Dhabi
Period8/12/1311/12/13

Keywords

  • Automated Assessment
  • Histogram based Thresholding
  • Hölder Exponent
  • Posterior Capsule Opacification
  • Texture

Fingerprint

Dive into the research topics of 'Automated image assessment of posterior capsule opacification using Hölder exponents'. Together they form a unique fingerprint.

Cite this