Accurate segmentation of ultrasound images using the motion cue

摘要

Accurate segmentation of ultrasound images is desired to measure tissue shapes and sizes. But in some cases, it is difficult to segment ultrasound images purely based on static intensity values. In this paper, we propose a novel segmentation framework by using the motion cue. We integrate a new speckle tracking algorithm into a level set framework to iteratively refine the segmentation and motion estimation at the boundary regions of two different tissues. The balance between local motion contrast of different tissue regions and the smoothness constraint will drive the active contour to the boundaries of different tissues. Experiments on both simulated data and phantom data show that our method has achieved better results than previous methods.