The application of adaptive filters for motion prediction in visually tracked laparoscopic surgery

Abstract

More and more research teams propose robotic systems to control the position of the laparoscope. In many of them the visual tracking principle is applied. However as experiments have shown tracked markers may often be obscured during a surgery. Within the article a study on the possibility of the application of a predictive algorithm to anticipate the position of a chirurgical tool is presented. This system would generate an estimate of the position whenever tracked tool’ s marker is obscured. As its basis the LMS algorithm was chosen, to which supplementary modifications were introduced. Experimental simulations prove that the algorithm greatly reduces the amount of missing samples during real time operation. Obtained predicted positions of the chirurgical tool reasonably approximate its true position.