Automatic 3D Model Creation during Minimally Invasive Surgery

From Master Projects
Jump to: navigation, search

has title::Automatic Model Creation during Minimally Invasive Surgery
status: finished
Master: project within::Internet and Web Technology
Student name: student name::Balázs Kőszegi
Start start date:=2011/12/01
End end date:=2012/05/31
Supervisor: Thilo Kielmann
Second reader: has second reader::Ben van Werkhoven
Company: has company::Personal Space Technologies
Thesis: has thesis::Media:Thesis.pdf
Poster: has poster::Media:Posternaam.pdf

Signature supervisor



Minimally Invasive Surgery (MIS) is a common surgical technique whereby patients are treated for their illness while inflicting minimal trauma. One of the procedures that is often performed using MIS is the removal of malignant tumors from the bladder. With this type of procedure the operator enters the bladder of the patient using an endoscope, locates and then removes one or more tumors using a small knife attached to the tip of the endoscope.

A particularly difficult situation can occur when multiple tumors are present. In this case the removal of one tumor can cause blood to flow and create a murky endoscope image. This makes it more difficult for a surgeon to find the remaining tumors and it raises the chance that tumors may be overlooked.

To assist the surgeon during this kind of procedure PS-Tech has developed an augmented reality in-patient navigation system. With this system the endoscope is tracked using an optical tracking system; at any time the exact location of the endoscope is known. This allows operators to locate and mark 3D positions within the bladder and find them back based on navigational instructions given by the system.

During the procedure the system collects a lot of interesting video and spatial data. Much of this data is currently not used to its full potential. The goal of this research project is to be able to exploit this information, such that the operator not only sees the current view from the endoscope but also has an instantaneous view of the entire bladder.

This project has the following expected outcome:

A software system that reconstructs a 2D panoramic model of the inside of the bladder, using the camera image stream during an endoscopic procedure along with estimates of camera positions of each image. The system selects camera images based on location and quality, such as sharpness, contrast, etc, corrects for distortion, lighting and color differences in the images, and provides a complete model of the bladder presented in a standardized fashion based on anatomical landmarks. It provides real-time feedback to the user on model construction.

Optionally, the system is extended to obtain a fully textured 3D model of the bladder.