Learning ultrasound scanning skills from human demonstrations

Recently, the robotic ultrasound system has become an emerging topic owing to widespread use of medical ultrasound. However, it is still a challenging task model and transfer skill from physician. In this paper, we propose learning-based framework acquire scanning skills human demonstrations. First, are encapsulated into high-dimensional multi-modal in terms interactions among images, probe pose contact force. The parameters learned using data collected skilled sonographers' Second, sampling-based strategy proposed with adjust extracorporeal process guide newbie sonographer or robot arm. Finally, robustness validated experiments on real sonographers.