Robotic-assisted Respiration-Corrected 4D Ultrasound Imaging for Image-guided Interventions

Widely used freehand 3D ultrasound systems do not provide corrections to effectively track breathing-induced moving organs. In this paper, a novel 4D ultrasound imaging method is proposed to create a sequence of breathing-corrected 3D ultrasound images from multiple 2D US image sequences, which is acquired by using a robotic arm to hold a 2D US probe and tilt it to regularly scan the liver. An image-based method is introduced to label the respiratory phases of the acquired images. Based on the extracted phases, the acquired 2D images are then resampled to reconstruct a sequence of respiration-corrected 3D ultrasound images. The reconstructed 3D images can capture the moving 3D liver during half a breathing cycle. Initial experiments demonstrate that the proposed method is feasible to visualize the 3D liver motion.