Retinal image segmentation using double-scale non-linear thresholding on vessel support regions

Abstract

Retinal vessel segmentation is a critical indicator of diagnosis, screening, and treatment of cardiovascular and ophthalmologic diseases. Due to the fact that the retinal vessels usually have some tiny structures and blurred boundaries, especially with remarkable noises, it is difficult to correctly segment the vascular networks. In this study, the authors propose a double-scale non-linear thresholding method based on vessel support regions. First, the double-scale filtering method is applied to enhance the contrast between the foreground vascular and the background stuffs. Second, they segment the fine and coarse vessels by the corresponding adaptive local thresholding and fixed-ratio thresholding method. Finally, they obtain the binary segmentation by fusion of fine and coarse vessels. Experiments are conducted on the publicly available DRIVE and STARE datasets, which show the effectiveness of the proposed method on retinal vessel segmentation.

Keywords

• blood vessels
• image segmentation
• eye
• medical image processing
• diseases
• image enhancement
• binary segmentation
• fine vessels
• coarse vessels
• retinal vessel segmentation
• retinal image segmentation
• vessel support regions
• cardiovascular diseases
• ophthalmologic diseases
• retinal vessels
• double-scale nonlinear thresholding method
• double-scale filtering method
• fixed-ratio thresholding method
• adaptive local thresholding
• double-scale non-linear thresholding
• contrast enhancement