Automatic Detection and Distinction of Retinal Vessel Bifurcations and Crossings in Colour Fundus Photography

Abstract

Abstract:The analysis of retinal blood vessels present in fundus images, and the addressing of problems such as blood clot location, is important to undertake accurate and appropriate treatment of the vessels. Such tasks are hampered by the challenge of accurately tracing back problems along vessels to their source. This is due to the unresolved issue of distinguishing automatically between vessel bifurcations and vessel crossings in colour fundus photographs. In this paper, we present a new technique for addressing this problem using a convolutional neural network approach to firstly locate vessel bifurcations and crossings and then to classifying them as either bifurcations or crossings. Our method achieves high accuracies for junction detection and classification on the DRIVE dataset and we show further validation on an unseen dataset from which no data has been used for training. Combined with work in automated segmentation, this method has the potential to facilitate: reconstruction of vessel topography, classification of veins and arteries and automated localisation of blood clots and other disease symptoms leading to improved management of eye disease.

A Multi-task Network to Detect Junctions in Retinal Vasculature

“…We also present a junction probability map, which significantly facilitates finding junctions by removing possible variations in vessel thickness. Experiments on the DRIVE and IOSTAR show our method outperformed Pratt et al's method [10], despite operating directly on fundus images.…”

“…One key advantage of this network is that -despite the ability to perform more than one task -it has relatively fewer parameters, slightly over one million, than those of many state-of-the-art networks. For example, Res16 over eleven million parameters, which was used by Pratt et al to classify image patches with junctions [10]. Also, pretraining can provide a good initialization for the network parameters and facilitate training of the network in the face of the scarcity of labeled data.…”

“…However, they have been rarely applied to junction detection in retinal vasculature. As far as we are aware, the only study is Pratt et al 's [10], where a popular network was trained as a classifier to identify image patches with junctions. The image patches contained vessel skeletons, generated from binary vessel maps.…”

Deep Learning in Cardiology

“…There are also other uses such as bifurcation/crossing identification. Pratt et al [191] trained a ResNet18 to identify small patches which include either bifurcation or crossing. Another ResNet18 was trained on patches that have been classified to have bifurcations and crossings to distinguish the type of vessel junction located.…”

Deep-Channel uses deep neural networks to detect single-molecule events from patch-clamp data

“…Convolutional neural network (CNN) layers are a powerful component of 62 deep learning useful for learning patterns within complex data. 2-dimensional (2D) CNNs are most 63 commonly applied to computer vision [19][20][21] and we have previously used them for automatic diagnosis 64 of retinal disease in images 22 . An adaptation of the 2D CNN is the one-dimensional (1D) CNN.…”