Ontology-based surgical assistance system for instruments recognition

Abstract

An ontology-based surgical assistance system was developed using ontology engineering, point cloud library and semantic technologies.An ontology, which provides knowledge base encoded with surgical instruments three dimensional model information as instances, was implemented.Colour-based region growing algorithm was used to segment the instruments and configured with the application.The final results were evaluated using the application interface, which allows to query on the ontology and to find an instrument's model that could be used for the instrument recognition and robotic assistance.We have segmented surgical instruments with approximately 64% accuracy and were able to integrate segmented models with the ontology.The results of model detection through ontology, for recognition, has sensitivity of 100%, specificity of 99.1% and accuracy 99.5%.