Information
- Publication Type: Master Thesis
- Workgroup(s)/Project(s): not specified
- Date: 2025
- TU Wien Library: AC17679808
- Open Access: yes
- First Supervisor: Renata Georgia Raidou
- Pages: 128
- Keywords: Uncertainty Propagation, Uncertainty, Visualization Pipeline, Medical Visualization Pipeline, Provenance, Parameter Sensitivity Analysis, Monte Carlo Methods
Abstract
Quantifying, raising awareness, and visualizing uncertainty stand as challenges in data visualization, especially in critical application domains such as medicine. Medical diagnosis and following treatment are always based on human decision-making, which itself is prone to uncertainty due to subjectivity or perception. Furthermore, decisions are often taken by analyzing measurements or images, which themselves are affected by uncertainty, caused by effects such as noise or resolution limitations. Thus, the whole process of diagnosis in clinical environments is concerned with interwoven uncertainties that accumulate and may change a pipeline's result substantially, potentially with detrimental effects on the patient's health, if uncertainties are not considered. This work aims to contribute to unraveling the complex interplay of uncertainties within the medical visualization pipeline. We do so by investigating the complex phenomena of uncertainty propagation in the medical visualization pipeline, in combination with extracting and analyzing provenance information from the pipeline encapsulated in an interactive framework. As a consequence, we utilize the provenance information, which can be seen as a complete history of the pipeline, to compare uncertainty propagation results of distinct pipeline states and thus gain insights into the behavior of uncertainty. In order to demonstrate the conceptual effectiveness of the framework, meaningful usage scenarios are presented. Those lay out simple and more complex scenarios to analyze the behavior and impact of different sorts of parameters present in the pipeline. Furthermore, we present ways in which a user can express their uncertainty for certain image regions or parameters and thereby gain insights into the impact of the specified uncertainties. The usage scenarios emphasize both positive and negative aspects of the framework and thus provide users with the means to assess the underlying work independently.
Additional Files and Images
Weblinks
BibTeX
@mastersthesis{haeusle-2025-uup,
title = "Unraveling uncertainty propagation in the medical
visualization pipeline",
author = "Gabriel H\"{a}usle",
year = "2025",
abstract = "Quantifying, raising awareness, and visualizing uncertainty
stand as challenges in data visualization, especially in
critical application domains such as medicine. Medical
diagnosis and following treatment are always based on human
decision-making, which itself is prone to uncertainty due to
subjectivity or perception. Furthermore, decisions are often
taken by analyzing measurements or images, which themselves
are affected by uncertainty, caused by effects such as noise
or resolution limitations. Thus, the whole process of
diagnosis in clinical environments is concerned with
interwoven uncertainties that accumulate and may change a
pipeline's result substantially, potentially with
detrimental effects on the patient's health, if
uncertainties are not considered. This work aims to
contribute to unraveling the complex interplay of
uncertainties within the medical visualization pipeline. We
do so by investigating the complex phenomena of uncertainty
propagation in the medical visualization pipeline, in
combination with extracting and analyzing provenance
information from the pipeline encapsulated in an interactive
framework. As a consequence, we utilize the provenance
information, which can be seen as a complete history of the
pipeline, to compare uncertainty propagation results of
distinct pipeline states and thus gain insights into the
behavior of uncertainty. In order to demonstrate the
conceptual effectiveness of the framework, meaningful usage
scenarios are presented. Those lay out simple and more
complex scenarios to analyze the behavior and impact of
different sorts of parameters present in the pipeline.
Furthermore, we present ways in which a user can express
their uncertainty for certain image regions or parameters
and thereby gain insights into the impact of the specified
uncertainties. The usage scenarios emphasize both positive
and negative aspects of the framework and thus provide users
with the means to assess the underlying work independently.",
pages = "128",
address = "Favoritenstrasse 9-11/E193-02, A-1040 Vienna, Austria",
school = "Research Unit of Computer Graphics, Institute of Visual
Computing and Human-Centered Technology, Faculty of
Informatics, TU Wien",
keywords = "Uncertainty Propagation, Uncertainty, Visualization
Pipeline, Medical Visualization Pipeline, Provenance,
Parameter Sensitivity Analysis, Monte Carlo Methods",
URL = "https://www.cg.tuwien.ac.at/research/publications/2025/haeusle-2025-uup/",
}
thesis