Quality Enhancement and Relation-Aware Exploration pipeline for Volume Visualization

The Hong Kong University of Science and Technology
Department of Computer Science and Engineering


PhD Thesis Defence


Title: "Quality Enhancement and Relation-Aware Exploration pipeline for
Volume Visualization"

By

Mr. Ming-Yuen Chan


Abstract

With the advance of data acquisition methods, various kinds of high 
quality volumetric datasets are now available for medical and scientific 
purposes. Due to the explosive increase in size and complexity of the 
data, the analysis tasks involved pose difficult problems. Volume 
visualization provides an effective solution which aims at delivering 
insights of the underlying details or concepts by the means of computer 
graphics for creating useful views on the data. In fact, effective 
visualization requires the support of intuitive and interactive 
exploration techniques as well as proper rendering and presentation of the 
data in order to facilitate the users in acquiring insights and useful 
information from the volumes. This thesis focuses on the improvement of 
the rendered image quality and the exploration pipeline to facilitate the 
visualization process.

In this thesis, we first present two exploration techniques for 
visualizing volumetric data. Instead of performing tedious manipulation on 
the volumes, we suggest a quality-driven camera path planning method and a 
relation-aware visualization pipeline to allow semi-automatic exploration 
of volumes. Volume features and high-level spatial relations are 
considered to determine the proper views on the volumes and our objective 
is to reveal this important information using the proposed visualization 
techniques. In the second part, the issues of rendered image quality are 
discussed and we propose two image quality enhancement approaches to 
ensure that the features in the volume are faithfully presented in the 
rendered images and correct perception of the volumes is obtained by 
viewers. An adaptive enhancement framework is also proposed to adjust the 
rendering parameters for obtaining optimal rendered images. To demonstrate 
the effectiveness of our solutions, experiments are conducted on different 
kinds of datasets and the results are evaluated and discussed.


Date:			Monday, 17 August 2009

Time:			10:00am-12:00noon

Venue:			Room 3402
 			Lifts 17-18

Chairman:		Prof. Zhiyu Yang (PHYS)

Committee Members:	Prof. Huamin Qu (Supervisor)
 			Prof. Albert Chung
 			Prof. Dit-Yan Yeung
 			Prof. Weichuan Yu (ECE)
 			Prof. Jian Huang (EECS, Univ. of Tennessee)


**** ALL are Welcome ****