Towards Automatic Robotic Ultrasound Imaging
Medical ultrasound (US) imaging is a well-established imaging tool and is employed for the diagnosis of countless diseases and pathologies in nearly every aspect of the human body. Traditional freehand-performed diagnostic US imaging has several drawbacks, including having inconsistencies in exam outcomes, being physically demanding, and increased disease transmission risk. US image acquisition using robotic ultrasound (RUS) has been extensively studied to address the above limitations. RUS systems leverage the dexterity and accuracy of the robotic arm to manipulate the US probe. Hence, direct contact between the patient and the clinician can be avoided. Additionally, probe maneuver can potentially be automated to enhance procedural repeatability. However, the development of an autonomous RUS system remains an open research challenge. In this project, we tackle this challenge by decomposing the autonomous imaging task into three sub-tasks: scanning target localization, optimal probe orientation control, and standardized imaging plane navigation using US image feedback. Furthermore, a novel framework is proposed to estimate organ pose from partial anatomical observation and a canonical organ template.
Advisor: Haichong K. Zhang (WPI)
Committee: Songbai Ji (WPI); Giovanni Pittiglio (WPI); and Junichi Tokuda (Brigham and Women's Hospital)
Zoom Link: https://wpi.zoom.us/j/2482368756