No-reference wavelet based retinal image quality assessment
Abdel Hamid LS, El-Rafei A, El-Ramly S, Michelson G, Hornegger J (2016)
Publication Status: Published
Publication Type: Conference contribution, Conference Contribution
Publication year: 2016
Publisher: CRC Press/Balkema
Pages Range: 123-130
ISBN: 9781138029262
URI: http://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84959286262&origin=inward
Abstract
Retinal Image Quality Assessment (RIQA) is an essential preliminarily step in Automatic Retinal Screening Systems (ARSS) to avoid misdiagnosis of retinal disease. In this work, a no-reference wavelet based RIQA algorithm suitable for real-time ARSS is introduced. Retinal images are classified into good and bad quality images using two proposed metrics. These measures are extracted from the wavelet decomposition subbands of the image to reflect sharpness and contrast. Wavelet level and regional analyses are performed to determine the most relevant levels and retinal region to be used in the RIQA. The algorithm is tested on two publicly available datasets with different resolutions. Area under the ROC curve (AUC) of 1 and 0.985 are obtained for the low and high resolution datasets respectively with runtime ranging between 0.1 and 2 seconds depending on image resolution. Comparison with other RIQA algorithms shows superiority of the proposed algorithm in both performance and time.
Authors with CRIS profile
Georg Michelson
Department of Ophthalmology
Joachim Hornegger
Lehrstuhl für Informatik 14 (Bild- und Sprachverarbeitung)
Involved external institutions
Egypt (EG)How to cite
APA:
Abdel Hamid, L.S., El-Rafei, A., El-Ramly, S., Michelson, G., & Hornegger, J. (2016). No-reference wavelet based retinal image quality assessment. In Proceedings of the 5th Eccomas Thematic Conference on Computational Vision and Medical Image Processing, VipIMAGE 2015 (pp. 123-130). CRC Press/Balkema.
MLA:
Abdel Hamid, L. S., et al. "No-reference wavelet based retinal image quality assessment." Proceedings of the 5th Eccomas Thematic Conference on Computational Vision and Medical Image Processing, VipIMAGE 2015 CRC Press/Balkema, 2016. 123-130.
BibTeX: Download