Automated segmentation of blood vessel in contrast enhanced plane wave ultrasound images

Abstract

Segmentation of blood vessel with ultrasound imaging has far reaching applications in medical diagnosis from deriving vascular mechanical properties to calculating wall shear stress (WSS). Contrast enhanced ultrasound (CEUS) provides valuable additional information and has a better tissue-to-lumen contrast, however, not much work has been done on segmentation in 3D CEUS images. The intra-lumen signal inhomogeneity and weak signal from tissue-vessel boundary in CEUS images create difficulties for existing methods. This study aims to develop a 3D segmentation algorithm that is more robust in tackling these problems in CEUS vessel segmentation. The result is evaluated both in-vitro and in-vivo.