Automatic Tortuosity-Based Retinopathy of Prematurity Screening System

Lassada SUKKAEW  Bunyarit UYYANONVARA  Stanislav S. MAKHANOV  Sarah BARMAN  Pannet PANGPUTHIPONG  

Publication
IEICE TRANSACTIONS on Information and Systems   Vol.E91-D   No.12   pp.2868-2874
Publication Date: 2008/12/01
Online ISSN: 1745-1361
DOI: 10.1093/ietisy/e91-d.12.2868
Print ISSN: 0916-8532
Type of Manuscript: PAPER
Category: Image Recognition, Computer Vision
Keyword: 
tortuosity,  retinopathy of prematurity,  segmentation,  

Full Text: PDF>>
Buy this Article




Summary: 
Retinopathy of Prematurity (ROP) is an infant disease characterized by increased dilation and tortuosity of the retinal blood vessels. Automatic tortuosity evaluation from retinal digital images is very useful to facilitate an ophthalmologist in the ROP screening and to prevent childhood blindness. This paper proposes a method to automatically classify the image into tortuous and non-tortuous. The process imitates expert ophthalmologists' screening by searching for clearly tortuous vessel segments. First, a skeleton of the retinal blood vessels is extracted from the original infant retinal image using a series of morphological operators. Next, we propose to partition the blood vessels recursively using an adaptive linear interpolation scheme. Finally, the tortuosity is calculated based on the curvature of the resulting vessel segments. The retinal images are then classified into two classes using segments characterized by the highest tortuosity. For an optimal set of training parameters the prediction is as high as 100%.