A Computer Vision Platform to Automatically Recognize Surgical Phases in Laparoscopic Pancreatoduodenectomy

Abstract

Introduction: Laparoscopic pancreatoduodenectomy (LPD) is a complex operation lasting several hours, whose intraoperative videos still require timely manual review and analysis. Computer vision(CV) is a kind of artificial intelligence(AI), which uses convolutional neural networks(CNN) to allow the computer to understand the images and videos. The purpose of our study is to develop a CV platform to recognize and extract surgical phases automatically in LPD videos.