Visual Analytics Applied to Image Analysis: From Segmentation to Classification

Rauber, P., 2017, [Groningen]: University of Groningen. 161 p.

Research output: ThesisThesis fully internal (DIV)Academic

Copy link to clipboard


  • Paulo Rauber
Image analysis is the field of study concerned with extracting information from images. This field is immensely important for commercial and scientific applications, from identifying people in photographs to recognizing diseases in medical images.

The goal behind the work presented in this thesis is providing mechanisms that allow humans to assist machines in image analysis tasks that are difficult to fully automate: image segmentation, feature selection, and image classification.

Image segmentation is the task of partitioning an image into objects of interest (e.g., identifying which pixels correspond to a person in an image). In this context, we propose a new technique that enables faster interactive segmentation and potentially richer feature extraction, which may lead to increased efficacy.

Image classification is the task of assigning a class label to an image based on generalization from examples (e.g., given images of a person, recognizing other images of this person). The traditional solution involves first representing each image by features (measurable characteristics) related to colors, textures, and shapes. In this context, we propose a new interactive visualization system that aims to provide insights that lead to the development of effective feature sets for image classification.

We also show how this system can be adapted to explore intermediary computational results of artificial neural networks, with the goal of enabling insights about how these networks operate, which again may lead to improvements along the image classification pipeline. This task also leads to the development of a new time-dependent visualization technique.
Original languageEnglish
QualificationDoctor of Philosophy
Awarding Institution
  • Falcão, A.X. (Alexandre), Supervisor, External person
  • Telea, Alexandru, Supervisor
  • de Rezende, P.J. (Pedro Jussieu), Supervisor, External person
  • Roerdink, Jos, Supervisor
  • Petkov, Nicolai, Assessment committee
  • Biehl, M. , Assessment committee
  • da Silva Torres, R. (Ricardo), Assessment committee, External person
  • Sminchisescu, Cristian, Assessment committee, External person
Award date20-Feb-2017
Place of Publication[Groningen]
Print ISBNs978-90-367-9288-2
Electronic ISBNs978-90-367-9287-5
Publication statusPublished - 2017

Download statistics

No data available

ID: 39606172