Retinal Vessel Detection in Retinopathy of Prematurity using Butterworth High-pass Filters and SVM
Niousha Hormozi1, Seyed Amirhassan Monadjemi2, Gholamali Naderian3
1Niousha Hormozi, Department of Computer Engineering, University of Isfahan, Isfahan Iran.
2Dr. Seyed Amirhassan Monadjemi, Department of Computer Engineering, University of Isfahan, Isfahan Iran.
3Dr. Gholamali Naderian, Department of Medical Science, Isfahan University, Isfahan Iran.
Manuscript received on 10 September 2013 | Revised Manuscript received on 19 September 2013 | Manuscript Published on 30 September 2013 | PP: 115-117 | Volume-3 Issue-4, September 2013 | Retrieval Number: D1164093413/13©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Retinopathy of prematurity (ROP) is an eye disease in premature infants. It mostly happens in babies weighing less than 0011 g and gestational age less than 10 weeks. The growth of retinal vessels are interrupted in premature infants and the retina has been unable to get enough oxygen and food. So, the delay in diagnosis may lead to blindness. Therefore it is necessary to follow the premature infants in regular checkups to assure that their vessel structure is growing normal. In this paper a high pass filter is used to track the retinal vessels and the energy criterion is computed for finding the percentage of area which is covered with blood vessels. The algorithm has been applied on 011 images including both mature and premature infants. The pictures are taken with a Retcam and labeled by an ophthalmologist. The result of this study was compared with ophthalmologist’s hand labels of diagnosis and it can detect the prematurity with a high specificity of 0111, sensitivity of %29.61 and accuracy of %59101
Keywords: Butterworth high-pass filter, ROP, SVM.
Scope of the Article: High Performance Computing