Segmentation of medical 3D volumes on a GPU

Anže Sodja (2013) Segmentation of medical 3D volumes on a GPU. EngD thesis.

Preview

Abstract

The main goal of the thesis is an efficient implementation of segmentation of medical 3D volumes obtained by means of 3D digital subtraction angiography. For the purposes of segmentation different algorithms are implemented – algorithms for data purification and algorithms for creating 3D models. To achieve better performance algorithms are implemented on a graphic process unit with OpenCL API. Complete segmentation process is integrated into a program designed for visualization of 3D models called Neck Veins. The thesis first describes all implemented algorithms, followed by a description of the library for general processing OpenCL. Finally, it describes an extension of the program Neck Veins. A comparison of the performance of certain algorithms on a central processing unit in the Java programming language, and the implementation on the graphics processing unit on the OpenCL API is made to show the effectiveness of the implementation.

Item Type:

Thesis (EngD thesis)

Keywords:

segmentation, GPU, algorithms, OpenCL, 3D volumes, Marching cubes

Number of Pages:

Language of Content:

Slovenian

Mentor / Comentors:

Name and Surname	ID	Function
doc. dr. Matija Marolt	271	Mentor

Link to COBISS:

http://www.cobiss.si/scripts/cobiss?command=search&base=50070&select=(ID=10204756)

Institution:

University of Ljubljana

Department:

Faculty of Computer and Information Science

Item ID:

2218

Date Deposited:

25 Sep 2013 15:31

Last Modified:

21 Oct 2013 10:00

URI:

http://eprints.fri.uni-lj.si/id/eprint/2218

Actions (login required)

View Item