Superpixel-Based Line Operator for Retinal Blood Vessel Segmentation
Automated detection of retinal blood vessels plays an important role in advancing the understanding of the mechanism, diagnosis and treatment of cardiovascular disease and many systemic diseases. Here, we propose a new framework for precisely segmenting vasculatures. The proposed framework consists of two steps. Inspired by the Retinex theory, a non-local total variation model is introduced to address the challenges posed by intensity inhomogeneities and relatively poor contrast. For better generalizability and segmentation performance, a superpixel based line operator is proposed as to distinguish between lines and the edges, and thus allows more tolerance in the position of the respective contours. The results on three public datasets show superior performance to its competitors, implying its potential for wider applications.
KeywordsVessel Segmentation Total variation Retinex Superpixel Line operator
This work was supported by National Science Foundation Program of China (61601029, 61602322), China Association for Science and Technology (2016QNRC001), and National Key Research and Development Program of China (2016YFB0401202).
- 22.Zhao, Y., Zhao, J., Yang, J., Liu, Y., Zhao, Y., Zheng, Y., Xia, L, Wang, Y.: Saliency driven vasculature segmentation with infinite perimeter active contour model. Neurocomputing (2017). http://dx.doi.org/10.1016/j.neucom.2016.07.077