Blood vessel classification into arteries and veins in retinal images

Abstract

The prevalence of diabetes is expected to increase dramatically in coming years; already today it accounts for a major proportion of the health care budget in many countries. Diabetic Retinopathy (DR), a micro vascular complication very often seen in diabetes patients, is the most common cause of visual loss in working age population of developed countries today. Since the possibility of slowing or even stopping the progress of this disease depends on the early detection of DR, an automatic analysis of fundus images would be of great help to the ophthalmologist due to the small size of the symptoms and the large number of patients. An important symptom for DR are abnormally wide veins leading to an unusually low ratio of the average diameter of arteries to veins (AVR). There are also other diseases like high blood pressure or diseases of the pancreas with one symptom being an abnormal AVR value. To determine it, a classification of vessels as arteries or veins is indispensable. As to our knowledge despite the importance there have only been two approaches to vessel classification yet. Therefore we propose an improved method. We compare two feature extraction methods and two classification methods based on support vector machines and neural networks. Given a hand-segmentation of vessels our approach achieves 95.32% correctly classified vessel pixels. This value decreases by 10% on average, if the result of a segmentation algorithm is used as basis for the classification.

Fundus Image Based Blood Flow Simulation of the Retinal Arteries

“…There are reports in the literature and active research trying to automatically classify arteries and veins, like from Kondermann et al [12], but this is not the focus of this study. The classification was done manually and the arteries were marked on the original image with red lines and the veins with blue lines.…”

Automatic Classification of Retinal Vessels Using Structural and Intensity Information

“…Vazquez et al [7] described a method which combines a color based clustering algorithm with a vessel tracking method. Kondermann et al [8] proposed two classification methods based on support vector machines and neural networks to classify retinal vessels, which are supported by features either extracted from vessel profiles or from a region of interest (ROI) defined around each vessel centerline point. Niemeijer et al [9] proposed an automatic method for A/V classification also using image features and a classifier.…”

Retinógrafo coaxial não-midriático

“…Como exemplo, hipertensão e aterosclerose causam alterações na razão entre os diâmetros das artérias e veias da retina, conhecida por razão arteríolo-venular (AVR). 49 Um decréscimo dessa razão, isto é, o afinamento das artérias e o alargamento das veias, está relacionado ao acréscimo do risco de derrame e infarto do miocárdio. 3,50,51 Todas essas doenças podem ser diagnosticadas através de retinógrafos, que realizam dois tipos de exames: retinografia e angiografia.…”

“…Artérias são mais claras porque transportam sangue com maior concentração de oxigênio, enquanto que veias transportam sangue com baixa concentração de oxigênio e, por isso, são mais escuras. 49 Nas Figuras 8.13 e 8.14 são exibidas retinografias do mesmo paciente, que foi solicitado a olhar um pouco para o lado de modo que o nervo óptico estivesse centralizado na imagem. Nessas imagens, os nervos ópticos são vistos com grande detalhe, mostrando a capacidade do RCNM de obter retinografias de alta qualidade.…”

“…Entretanto, a distinção entre os dois tipos de vascularização é determinante para o correto diagnóstico. 49 Embora já existam métodos para esse tipo de medida, é possível utilizar o RCNM para melhorar o diagnóstico, baseado nas diferenças dos espectros de absorção dos sangues com alta e baixa concentração de oxigênio. A medida da concentração de oxigênio no corpo humano já é realizada com base nessa diferença de absorção para os comprimentos de onda 660 e 940 nm.…”