Neural Network based system for real-time organ recognition by analysis of sequence of endoscopic images received during surgical operation

Abstract

This paper designs two Neural Network (NN) based systems for distinguishing and real-time recognition of internal organs on sequence of endoscopic images during abdominal surgery. First NN-based system proposed in this paper is designed for recognition of several different internal organs on color endoscopic images. Second NN-based system is designed for real-time recognition of presence of a particular internal organ on a sequence of color images (video stream) from endoscope. Restricted connectivity structure of the network makes possible decomposition of the image during the analysis and significantly reduces the number of parameters thus making training easier, faster and more accurate. The algorithms proposed in the paper are implemented in software application and their effectiveness is demonstrated on simulations.