Autonomous Robotic Carotid Ultrasound Examination

Abstract

Ultrasound (US) plays an essential role in the diagnosis of carotid artery diseases. Nowadays, most of the US examinations are done in a free-hand way. In order to lift the physical burden of doctors and increase the accuracy and reproducibility of the US acquisition, robotic arms offer a perfect alternative solution. Due to the superior performance in accuracy and repeatability, robotic technologies have been employed to develop a robotic US system (RUSS) to overcome the limitation of operator-variation and further improve the clinical acceptance of US modality. Aside from robotic technologies, thanks to the rapid development in the field of deep neural networks, it is feasible to have an autonomous segmentation network to provide doctors with standardized diagnosis images or even medical reports. Possible topics to offer could be:

• Segmentation of carotid plaques with deep neural networks.
• Autonomous localization of the plaque area and standard US plane navigation.
• Trajectory planning for US sweeps.
• Carotid plaque characteristics based on deep neural networks.
• Automatic medical report generation

Keywords: Ultrasound imaging segmentation, Robotic ultrasound, Standard planes navigation.