Robotic ultrasound scanning for deep venous thrombosis detection using RGB-D sensor

Abstract

This paper proposes a novel robotic US scanning system to scan the lower limbs of the patient for Deep Venous Thrombosis (DVT) detection. The system is composed of a robotic arm, a linear US probe with marker, and an RGB-D as a vision servo. The robot follows the navigation of the RGB-D sensor and scans the legs automatically. Kinect is chosen here to obtain the 3D images of legs and probe. The legs are segmented in RGB images and depth images real-time. The probe's position and orientation are registered according to the marker. The distance between the probe and the scanned leg is calculated every frame for safely scanning without force sensor. Both phantom and human leg platforms with robot and US probe and Kinect were built up and scan experiments were carried out. The transformation matrixes between two legs, depth imaging, Kinect and US probe coordinate systems, were obtained. The distances between probe and leg was calculated. Results from the initial experiment indicate the idea is feasible and promising greatly by improving the inspection efficiently. Clinically, the method can be implemented for pre-operative procedures to predict the risk of DVT, and this may improve the US scanning efficacy.