Read/Search this Article
Abstract
眼底画像を解析することによって,高血圧性網膜症の所見である細動脈狭窄を定量的に評価し,眼科医の診断を支援するコンピュータ支援診断システムを開発することが本研究の目的である.本研究では,Black top hat変換と2重リングフィルタを組み合わせて血管抽出し,血管領域内のR,G,B成分を用いて線形判別関数を求めることによって主幹動静脈の分類を行なった.その結果,主幹動静脈の分類精度は88.2%であった.さらに,分類した血管から,細動脈狭窄の診断に必要な並走する動静脈に絞り,動静脈の血管径を計測することによってA/V比を算出した.その結果,平均誤差と標準偏差は0.06±0.05であった.
The purpose of this study is to develop a computer-aided diagnosis(CAD)system for early detection of arteriolar narrowing on retinal images. In this study, blood vessels were detected using the black top hat transformation and a double ring filter, and major arteries and veins were classified by linear discriminant analysis using RGB components of blood vessel regions. As a result, the classification accuracy of major arteries and veins was 88.2%. In addition, major arteries and veins that run parallel to each other were selected for evaluation of arteriolar narrowing, and A/V(artery and vein)ratios were calculated based on their diameters. The average error±standard deviation of the proposed method was 0.06±0.05.
Journal
- IEICE technical report. [List of Volumes]
-
IEICE technical report. 109(407), 189-193, 2010-01-21 [Table of Contents]
The Institute of Electronics, Information and Communication Engineers