Activity recognition applied to the organ removal in surgical robotic system

Bibliographic Information

Other Title
  • 遠隔手術支援システムにおける臓器摘出の動作認識に関する研究

Abstract

<p>Surgical robotic systems enable doctors to perform surgery on a patient at a remote location. In telesurgery, several assistance methods have been developed to overcome the problem of communication delay because of long distance. In order to automate surgical tasks, real-time recognition of surgical activity is indispensable. In this study, we designed a network model based on temporal convolutional network and MobileNet for activity recognition. This proposed model achieved 57% in accuracy and 0.42 in F-measure. We also propose a model using segmentation information, and it improved the accuracy by 4% and F measure by 0.09.</p>

Journal

Details 詳細情報について

Report a problem

Back to top