Joint Learning of Vessel Segmentation and Artery/Vein Classification with Post-processing
This work addresses the time-consuming task of reading retinal images for disease diagnosis, though it appears incremental as it builds on existing automated methods.
The paper tackles the problem of automated retinal image analysis by proposing a two-step method for vessel segmentation and artery/vein classification, achieving an AUC of 0.98 for segmentation and an accuracy of 0.92 in classification on the DRIVE dataset.
Retinal imaging serves as a valuable tool for diagnosis of various diseases. However, reading retinal images is a difficult and time-consuming task even for experienced specialists. The fundamental step towards automated retinal image analysis is vessel segmentation and artery/vein classification, which provide various information on potential disorders. To improve the performance of the existing automated methods for retinal image analysis, we propose a two-step vessel classification. We adopt a UNet-based model, SeqNet, to accurately segment vessels from the background and make prediction on the vessel type. Our model does segmentation and classification sequentially, which alleviates the problem of label distribution bias and facilitates training. To further refine classification results, we post-process them considering the structural information among vessels to propagate highly confident prediction to surrounding vessels. Our experiments show that our method improves AUC to 0.98 for segmentation and the accuracy to 0.92 in classification over DRIVE dataset.