diplomsko delo
Žiga Lesar (Author), Matija Marolt (Mentor)

Abstract

V delu obravnavamo postopke za vizualizacijo medicinskih volumetričnih podatkov v realnem času, ki bi lahko povečali hitrost in zanesljivost diagnostičnih procesov. V ta namen smo razvili in primerjali tri vizualizacijske metode, ki temeljijo na algoritmu metanja žarkov: izris nivojskih ploskev, projekcijo največje intenzitete in emisijsko-absorpcijski model. Dodali smo tudi vizualna efekta ambientnega zastiranja in globinske ostrine. Vse metode smo optimizirali za čim hitrejše izvajanje na grafičnih procesorjih. Pri tem smo predstavili dve inovativni rešitvi: adaptivno vzorčenje in adaptivno rekonstrukcijo, ki pripomoreta k večji hitrosti izrisa. Opravili smo performančno analizo vseh uporabljenih metod, ki pokaže na učinkovitost predstavljenih pohitritev ter na ozka grla pri implementaciji. Izvedli smo tudi uporabniško evalvacijo, v kateri so specialisti s področja radiologije ovrednotili praktično primernost vizualizacijskih metod.

Keywords

vizualizacija medicinskih podatkov;angiografija;metanje žarkov;nivojske ploskve;emisijsko-absorpcijski model;OpenCL;računalništvo;računalništvo in informatika;univerzitetni študij;diplomske naloge;

Data

Language: Slovenian
Year of publishing:
Typology: 2.11 - Undergraduate Thesis
Organization: UL FRI - Faculty of Computer and Information Science
Publisher: [Ž. Lesar]
UDC: 004:61(043.2)
COBISS: 10773844 Link will open in a new window
Views: 0
Downloads: 0
Average score: 0 (0 votes)
Metadata: JSON JSON-RDF JSON-LD TURTLE N-TRIPLES XML RDFA MICRODATA DC-XML DC-RDF RDF

Other data

Secondary language: English
Secondary title: Real-time visualization of medical volumetric datasets
Secondary abstract: Our work discusses methods for real-time visualization of medical volumetric data sets, which can lead to faster and more reliable diagnostic processes. We have developed three visualization methods based on the ray casting algorithm: isosurface rendering, maximum intensity projection, and the emission-absorption model. We introduced two visual effects that enhance the final rendering: ambient occlusion and depth of field. We optimized the developed visualization methods for fast execution on graphics processing units and in the process introduced two novel methods: adaptive sampling and adaptive reconstruction, which contribute to faster rendering. We conducted a performance evaluation of the methods, which showed the efficiency of the implemented speed-ups and revealed the bottlenecks in the implementation. We also carried out a user evaluation study, where radiology specialists evaluated the practical suitability of the visualization methods.
Secondary keywords: visualization of medical data;angiography;ray casting;isosurfaces;emission-absorption model;OpenCL;healthcare applications;computer science;computer and information science;diploma;
File type: application/pdf
Type (COBISS): Bachelor thesis/paper
Study programme: 1000468
Thesis comment: Univerza v Ljubljani, Fak. za računalništvo in informatiko
Pages: 78 str.
ID: 8739367