Figure 1: Example of Boundary-Aware Network for Medical Image Segmentation.
Abstract
Although numerous deep learning models have achieved remarkable success in many medical image segmentation tasks, accurate segmentation of organs or tumors remains challenging, due to the varying sizes of organs and tumors and the ambiguous boundaries among them. In this paper, we propose a boundary-aware network (BA-Net) for medical image segmentation. This model contains a shared encoder, a boundary decoder, and a segmentation decoder. The multi-scale deep supervision strategy is adopted on both decoders, which can alleviate the issues caused by variable target sizes. The boundary probability maps produced by the boundary decoder at each scale are used as attention to enhance the segmentation feature maps. We evaluated the BA-Net on the Abdominal Multi-Organ Segmentation (AMOS) Challenge dataset and achieved an average Dice score of 89.29$\%$ for multi-organ segmentation on CT scans and an average Dice score of 71.92$\%$ on MRI scans. Also, we evaluated the BA-Net on the Kidney PArsing (KiPA) Challenge dataset and achieved an average Dice score of 89.65$\%$ for kidney structure segmentation on CTA scans using 4-fold cross-validation. The results demonstrate the effectiveness of the BA-Net.