Voice based Segmentation of Laparoscopic Surgical Tools and its Image Enhancement

Abstract

With more and more people preferring Laparo-scopic surgeries because of its benefits like smaller scars and being relatively painless in nature, there is immense need to solve the issues the doctors face while performing a Laparoscopic surgery. In this paper we propose a surgical tool segmentation and image enhancement method which can be controlled with a voice based interface. The proposed method solves problems like glare and fogging in addition to segmenting the surgical tools. A voice based interface was used so that the doctor does not have to get distracted while performing the surgery. This method can be helpful to surgeons and can also be helpful in achieving automated surgeries. For image segmentation we use Mask-RCNN and Kaldi Voice Recognition toolkit for the voice based interface. Kaldi toolkit has a great influence on the time taken by the system to understand the query.