Home | deutsch  | Legals | Data Protection | Sitemap | KIT

Automatic Surgical Tool Recognition in Cataract Surgery Videos using Machine Learning

Automatic Surgical Tool Recognition in Cataract Surgery Videos using Machine Learning
type:Bachelor thesis
tutor:

Dipl. Inf. Andreas Wachter

Download as PDF

Motivation and Background
Nowadays cataract surgery is a very often performed surgery. During a cataract surgery, he natural lens of the eye is replaced by a new artificial lens. The main surgical workflow has no large variation. So, this surgery procedure is very attrac- tive for automatic recognition of e.g. surgical tools or detection of the phase within the surgical workflow with potential applications. For analysis, medical images or videos would be used as input for a machine learning algorithm in most cases.

Task
The goal of the thesis is to indicate, which tools are being used by the surgeon at each instant. This goal could be achieved by using neuronal networks (NN). In a first step the topology of the NN must be characterized and defined. For this step, a pretrained NN can be selected. The second step is to implement and train the NN.
The main motivation for annotating tool usage is to design efficient solutions for surgical workflow analysis, with potential applications in report generation, surgical training and real-time decision support.