Out-of-Plane Motion Detection System Using Convolutional Neural Network for US-guided Radiofrequency Ablation Therapy

Abstract

Radiofrequency ablation therapy support system aims to display the true tumor position by tracking the tumor at the time of treatment. As a major cause of the error, the tumor moves in the direction which is perpendicular to the scan plane of the ultrasound image. Therefore, in the proposed method, the six-axis movement amount is estimated from two ultrasound images by a convolution neural network. By this method, we estimated the amount of movement in the direction perpendicular to the ultrasound image plane and aimed at tracking more accurately. We conducted an experiment to estimate the probe movement amount with respect to the liver phantom and confirmed the certain effectiveness of this method.