Ensemble of Deep Convolutional Neural Networks for Learning to Detect Retinal Vessels in Fundus Images

Vision impairment due to pathological damage of the retina can largely be through periodic screening using fundus color imaging. However the with large scale screening is the inability to exhaustively detect blood vessels crucial to disease diagnosis. In this work we present a imaging framework using deep and ensemble learning for reliable of blood vessels in fundus color images. An ensemble of deep neural networks is trained to segment vessel and non-vessel areas a color fundus image. During inference, the responses of the individual of the ensemble are averaged to form the final segmentation. In evaluation with the DRIVE database, we achieve the objective of detection with maximum average accuracy of 94. 7\\% and area under ROC of 0. 9283.