Title: Real-Time Instrument Segmentation in Robotic Surgery using Auxiliary Supervised Deep Adversarial Learning

Abstract: Robot-assisted surgery is an emerging technology which has undergone rapid growth with the development of robotics and imaging systems. Innovations in vision, haptics and accurate movements of robot arms have enabled surgeons to perform precise minimally invasive surgeries. Real-time semantic segmentation of the robotic instruments and tissues is a crucial step in robot-assisted surgery. Accurate and efficient segmentation of the surgical scene not only aids in the identification and tracking of instruments but also provided contextual information about the different tissues and instruments being operated with. For this purpose, we have developed a light-weight cascaded convolutional neural network (CNN) to segment the surgical instruments from high-resolution videos obtained from a commercial robotic system. We propose a multi-resolution feature fusion module (MFF) to fuse the feature maps of different dimensions and channels from the auxiliary and main branch. We also introduce a novel way of combining auxiliary loss and adversarial loss to regularize the segmentation model. Auxiliary loss helps the model to learn low-resolution features, and adversarial loss improves the segmentation prediction by learning higher order structural information. The model also consists of a light-weight spatial pyramid pooling (SPP) unit to aggregate rich contextual information in the intermediate stage. We show that our model surpasses existing algorithms for pixel-wise segmentation of surgical instruments in both prediction accuracy and segmentation time of high-resolution videos.