Perfusion quantification of contrast-enhanced ultrasound images based on coherence enhancing diffusion and competitive clustering

The aim of this paper is to introduce a novel approach to quantify liver arterial perfusion. In this study, the contrast-enhanced ultrasound modality is used to investigate liver diseases. The results show the pertinence of this approach in liver carcinoma. The method is based on perfusion estimation, in a regional basis and characterized by a fully-automated tracking of lesion in images sequences based on image statistics. This approach proceeds in two steps in a new framework, a first PDE algorithm and a second segmentation step by competitive clustering.