Image Processing under Uncertainty

Technische Universität Kaiserslautern

Language:

Christina Gillmann

Dr Christina Gillmann finished her PhD in 2018 as part of the international research training group "Physical Modeling for Virtual Manufacturing Systems and Processes" at the University of Kaiserslautern. Her research is centered around the applicability of image processing techniques in different domain. Her projects origin from collaborations with a variety of institutions from Germany, the US and Colombia.

Let's have a?

Expertise

  • Image Processing
  • Uncertainty Visualisation
  • Machine Learning

Of interest to

  • Researcher and developer in Image Analysis
  • Researcher and developer in Uncertainty Analysis
Dawid Zawila/Unsplash
Christina Gillmann

Dr Christina Gillmann finished her PhD in 2018 as part of the international research training group "Physical Modeling for Virtual Manufacturing Systems and Processes" at the University of Kaiserslautern. Her research is centered around the applicability of image processing techniques in different domain. Her projects origin from collaborations with a variety of institutions from Germany, the US and Colombia.

Let's have a?

Expertise

  • Image Processing
  • Uncertainty Visualisation
  • Machine Learning

Of interest to

  • Researcher and developer in Image Analysis
  • Researcher and developer in Uncertainty Analysis

Interview

Arthur Höring
Editor

How would you explain "uncertainty" to a non-expert?

Christina Gillmann
is typing…
Arthur Höring
Redakteur

How would you explain "uncertainty" to a non-expert?

Christina Gillmann
Doktorandin

Imagine you take a photo using the camera on your mobile phone. It comes out blurry. You are certain of that because you can see the photographed object before your eyes. But if you take an image from the medical field, for example, the MRT of a head you cannot be so sure if it is blurry or not. This issue explains what is meant by the uncertainty of images.

Arthur Höring
Redakteur

Your research examines image processing in many different areas. How much expertise did you need to acquire for each area and how easy or difficult was that for you?

Christina Gillmann
Doktorandin

To understand the data and the problems of a particular field it is crucial to know something about it. For the medical sector that was fairly easy for me since I had gone to several lectures and worked on that topic for over ten years. In other fields in which I had less experience, it took a bit longer to get used to it. There often the users themselves helped me by explaining the issues they had come across.

Arthur Höring
Redakteur

Towards the end of your thesis, you write that users need a higher acceptance towards new methods of image processing. How would you describe the relationship between user and software from your point of view?

Christina Gillmann
Doktorandin

In many areas, users view new software with scepticism. This is understandable because they know that it usually takes time and experience to be able to solve a problem.
In my opinion, many software solutions make the mistake of showing a too patronising attitude towards their users or they communicate results poorly or they are simply not user-friendly. If that changes, a new acceptance towards software solutions could be achieved.

Keywords

Image Processing, Uncertainty Visualisation

Summary

Novel image processing techniques have been in development for decades, but most of these techniques are barely used in real world applications. This results in a gap between image processing research and real-world applications; this thesis aims to close this gap. In an initial study, the quantification, propagation, and communication of uncertainty were determined to be key features in gaining acceptance for new image processing techniques in applications. This thesis presents a holistic approach based on a novel image processing pipeline, capable of quantifying, propagating, and communicating image uncertainty. This work provides an improved image data transformation paradigm, extending image data using a flexible, high-dimensional uncertainty model. Based on this, a completely redesigned image processing pipeline is presented. In this pipeline, each step respects and preserves the underlying image uncertainty, allowing image uncertainty quantification, image pre-processing, image segmentation, and geometry extraction. This is communicated by utilizing meaningful visualization methodologies throughout each computational step. The presented methods are examined qualitatively by comparing to the State-of-the-Art, in addition to user evaluation in different domains. To show the applicability of the presented approach to real world scenarios, this thesis demonstrates domain-specific problems and the successful implementation of the presented techniques in these domains.

Suggested citation

Gillmann, Christina. Image Processing under Uncertainty. Technische Universität Kaiserslautern, 2019, http://nbn-resolving.de/urn:nbn:de:hbz:386-kluedo-54707.

Repository

kluedo.ub.uni-kl.de

Identifiers

urn:nbn:de:hbz:386-kluedo-54707