Technical Reports


  • QI-2007-01
    I.W.O. Serlie, Lesion Conspicuity and efficiency of CT colonography with electronic cleansing based on a three-material transition model, 42 p. This report lists polyps that have been uncovered by electronic cleansing. Patients are included from public study data from the Walter Reed Army Medical Center (WRAMC) made available by the National Cancer Institute ( ). In addition to the CT volumes, colonoscopy reports are provided to provide the reference. The ‘3-D unfolded cube’ display has been chosen to display the colon surface in a more optimal manner than provided by forward and backward viewing directions, especially in the haustrations of the colon (1)(2).


  • QI-2006-01
    I.W.O. Serlie, F.M. Vos, H.W. Venema, L.J. van Vliet, CT Imaging Characteristics, 19 p, 4.512 Mb. This document explores basic aspects of CT imaging for Colonography: the noise characteristics, the spatial resolution and the model that describes the CT value across a material transition [1]. It is important to notice that the image processing discussed in this chapter is restricted to the edges of the colon. When applied to other areas or sources of data, the assumptions that are made have to be tested again.


  • QI-2004-01
    Michael van Ginkel, Cris L. Luengo Hendriks and Lucas J. van Vliet, A short Introduction to the Radon and Hough transforms and how they relate to each other, 9 p, 201 Kb. Extraction of primitives, such as lines, edges and curves, is often a key step in an image analysis procedure. The most popular technique for curve detection is based on the Hough transform. The original formulation of the Hough transform is inherently discrete. It is therefore difficult to assess which properties are inherent to the transform-based technique and which are due to its discrete nature. As other authors have pointed out before, the Hough transform is closely related to the Radon transform, in fact equivalent, if one is not too pedantic about the original formulations of the two transforms. With this report we hope to once again stress this relationship. The Radon transform formalism has two advantages over the Hough formalism. It has a well-founded mathematical basis and, in our opinion, is more intuitive as well.


  • PH-2003-01
    Pavel Paclík, Robert .P.W. Duin, Geert .M.P. van Kempen and Reinhardt Kohlus, Segmentation of multi-spectral images using the combined classifier approach, 10 p, 690 Kb. Appeared in: Image and Vision Computing, vol. 21, no. 6, pp. 473-482. Segmentation methods, combining spectral and spatial information, are essential for analysis of multi-spectral images. In this article, we propose such a method based on statistical pattern recognition algorithms and a combined classifier approach. A set of experiments is presented with multi-spectral images of detergent laundry powders acquired by imaging cross-sections with scanning electron microscopy using energy-dispersive X-ray microanalysis (SEM/EDX). The algorithm stability and the segmentation quality are investigated. The use of apriori information for the segmentation of images with similar spectral properties is studied as well. Finally, a comparison with probabilistic relaxation method for multi-spectral image segmentation is made.
  • PH-2003-02
    Cris L. Luengo Hendriks, Michael van Ginkel and Lucas J. van Vliet, Underestimation of the radius in the Radon transform for circles and spheres, 8 p, 140 Kb. In this technical report we compute the underestimation of the radius in the Radon transform for circles and spheres. Our implementation of the Radon transform uses spheres with a Gaussian profile, and normalises the grey-value of each of the spheres so that a very large sphere matching only a couple of segments will not get a higher confidence (value of the peak in the parameter space) than a very small sphere completely matched in the image. This normalisation causes an underestimation of the radius.
  • PH-2003-03
    Jurjen Caarls, Geometric Algebra with Quaternions, 3 p, 124 Kb. Quaternions are conveniant to store rotations because there are no singularities as in the case of euler angles. Furthermore, because there is no flip-over from 2ϖ to 0, a derivative can be defined as well. I found that people often give formulas involving quaternions, but not give the definitions they use. A number of different notations are in use in literature, so I will give here the definitions I use in my papers.


  • PH-2002-01
    Dick de Ridder and Robert P.W. Duin, Locally linear embedding for classification, 15 p, 2.1 Mb. Locally linear embedding (LLE) is a recently proposed unsupervised procedure for mapping high-dimensional data nonlinearly to a lower-dimensional space. In this paper, a supervised variation on LLE is proposed. This mapping, when combined with simple classifiers such as the nearest mean classifier, is shown to yield remarkably good classification results in experiments. Furthermore, a number of algorithmic improvements are proposed which should ease application of both traditional and supervised LLE by eliminating the need for setting some of the parameters.
  • PH-2002-02
    Bernd Rieger and Lucas J. van Vliet, A systematic approach to nD orientation representation, 11 p, 280 Kb. In this paper we present new insights in methods to solve the orientation representation problem in arbitrary dimensions. The gradient structure tensor is one of the most used descriptors for local structure in multi-dimensional images.We will relate its properties to the double angle method (2D) and the Knutsson mapping.We present a general scheme to reduce the dimensionality of the mappings needed to solve the orientation representation problem and derive some properties of these reduced mappings.



Name author: Webmaster ImPhys
© 2015 TU Delft