Force-based Automatic Identification and Skill Assessment of Forceps Manipulations

Haptic information is crucial in the execution of precise and dexterous manipulations. This paper addresses the challenge of skill evaluation during laparoscopic surgery by ( 1 ) automatic classification of basic manipulations, ( 2 ) evaluation of similarity between professional doctors and novices by a DP matching technique. First, manipulations with forceps during a laparoscopic surgery were categorized into four basic manipulations based on a video observation. Then, we developed forceps with a force sensing capability, and the basic manipulations were automatically classified based on the measured force information. By comparing with professional doctorsʼ basic manipulations database, the difference of manipulation skills between novices and doctors was evaluated by the DP matching method. We conducted ( I ) a tape peeling task and ( II) a tissue cutting task with novices and doctors. The experimental results show that the manipulations done by the doctors were much closer to the manipulations stored in the database than those done by the novices (p=0.0013 for the tape peeling task, p=0.021 for the tissue cutting task).