Towards Automated Initial Probe Placement in Transthoracic Teleultrasound Using Human Mesh and Skeleton Recovery
Year: 2026
Type: preprint
Abstract: Cardiac and lung ultrasound require operators to locate patient-specific intercostal acoustic windows and adjust probe position, rotation, and force to acquire standard views. In teleultrasound, experts may navigate using ultrasound images after approximate placement, but limited 3D perception makes remote guidance of initial placement difficult. We present tUS-PIPG, a framework for teleultrasound Patient registration and anatomy-informed Initial Probe placement Guidance. A novice captures calibrated RGB and depth images using a mixed reality headset. An edge server reconstructs the patient’s body surface and estimates skeletal landmarks from multiview RGB images and an accumulated depth point cloud. The method leverages spatial and temporal consistency across observations to achieve robust, training-free human registration. Skeletal landmarks guide estimation of intercostal regions, and the headset overlays virtual probe placement on the reconstructed body surface. We evaluated registration and placement guidance in a healthcare setting across multiple transthoracic cardiac and lung scan planes in situ. In a pilot study with five healthy volunteers, mean surface error was 15 mm, and mean error relative to palpated anatomical landmarks was 41.0 mm. These results suggest that the framework can effectively provide novice operators with a warm start in positioning the probe for remote expert adjustment.
Source: Authorea
Institution: University of British Columbia
Language: English
Field: Medicine
Domain: Health Sciences
Open Access status: green
OpenAlex ID: W7135232893
