A recent article by Van Hecke et al. [3] describes a framework for creating simulated diffusion-weighted images (DWI). The methodology allows for modeling intersubject variability, regional pathology, and noise and is quite useful for evaluation purposes. This article describes our ITK implementation of this work which we provide to the community as open source.

Simulated Diffusion-Weighted Imaging for the ITK Masses

In this article, a spline driven image slicer algorithm is presented. It is concretized through a vtkAlgorithm-inherited class that takes two inputs and gives two outputs. The first input is the volume from which the slice should reconstructed while the second one is the spline which tangent vector is used as slicing plane normal. The first output is a 2D image resliced from the volume.
The second output gives the geometric information of the slice plane location in the volume space. An example of use is given through a simulation of a dental panoramic view from a MDCT volume.

A Spline-Driven Image Slicer

The Hausdorff distance is a measure of the distance between sets of points. There are many advantages to using this metric compared to other similarity measures. This document describes a VTK class for computing the Hausdorff Distance between two sets of points. The main contribution, compared to other implementations, lies in the definition of the distance not only to the closest point but to the closest point in the represented surface, which yields an accurate measure even between undersampled surfaces. This is achieved by implementing a point-to-cell distance instead of a point-to-point. Furthermore, a plugin for ParaView was implemented, which is also available with the code. After introducing the interest of this distance, the VTK code is explained and illustrated with some examples.

A VTK Algorithm for the Computation of the Hausdorff Distance

A contribution to the Insight Toolkit intended to support the process of statistical
analysis in Computational Anatomy. The methods included here operate on open or closed triagulated surfaces (represented by a QuadEdgeMesh). The filter assigns basis function values as Point Data on each vertex of the Mesh.
This paper is accompanied with the source code, input data, parameters and output data that we used for validating the algorithm described in this paper. This adheres to the fundamental principle that scientific publications must facilitate reproducibility of the reported results.

Laplace Beltrami Filter on QuadEdge Meshes

Hilbert curves look pretty sweet which is why I wrote this class. Take a look at the pictures and see for yourself.

Why do walks with Hilbert seem to take so long?

Non-destructive evaluation(NDE) is a group of analysis techniques used in industry to evaluate the property of a material, component or product without causing damage. Among all the techniques, radiology and thermography are two most common methods that are used to inspect the property of the interior of product and combining the knowledge contained in images of two modalites requires the space alignment. In this article, a 2D/3D multi-modality image registration scheme is presented to find the alignment between thermal infrared(IR) image and CT image of a engine blade. 

2D/3D Multi-Modality Image Registration in Non-Destructive Evaluation

This document describes the implementation in ITK of the "point in circle" geometrical predicate.
Based on Jonathan Shewchuk's work which implements an exact version of the predicate using standard
floating point types and arithmetic, the implementation consist of an ITK wrapper around the 
public domain C routines made available by the author of the precedent paper.
Wrapper using itk::PointSet, itk:CellInterface and itk:Mesh / itk:QuadEdgeMesh APIs are provided
along with corresponding examples which should provide enough details for users to directly
copy paste code in their application.

The application in mind for us is an exact and robust implementation of a delaunay triangulation /
voronoi tesselation in ITK, and will be presented in a separate paper.

Exact Geometrical Predicate: Point in circle

This document presents a new class to apply an itkImageFunction to every pixel in an itkImage. This functionality is almost identical to itkUnaryFunctorImageFilter, but it uses an itkImageFunction rather than an itk::Functor. As some functionalities in ITK seem to have been implemented as image functors (itk::Functor::AND, itk::Functor::Atan, etc.) while other seems to have been implemented as itkImageFunction's (itk::BinaryThresholdImageFunction, itk::CentralDifferenceImageFunction, etc.), it seems reasonable to be able to apply any of these operations to an entire image.

The code is available here: http://review.source.kitware.com/#change,2008

ItkImageFunctionImageFilter: A New Filter To Apply An itkImageFunction To Every Pixel In An itkImage

In this document we present the implementation of three fuzzy clustering algorithms using the Insight Toolkit ITK. Firstly, we developed the conventional Fuzzy C-Means that will serve as the basis for the rest of the proposed algorithms. The next algorithms are the FCM with spatial constraints based on kernel-induced distance and the Modified Spatial Kernelized Fuzzy C-Means. Both of these introduce a Kernel function, replacing the Euclidean distance of the FCM, and spatial information into the membership function.

These algorithms have been implemented in a threaded version to take advantage of the multicore processors. Moreover, providing an useful implementation make it possible that classes work with 2D/3D images, different kernels and spatial shapes.

We included the source code as well as different 2D/3D examples, using several input parameters for the algorithms and obtaining the results generated on 2D/3D CT lung studies.

Fuzzy Clustering Algorithms for Image Segmentation

We have implemented filters for computing phase symmetry (PS) in ITK using steerable filters computed and applied in the Fourier domain. In some recent publications, phase symmetry has been shown to be a useful pre-processing measure for performing image registration and segmentation. Matlab implementations of the code are currently available, but may be slower than C++ versions and most importantly, do not work with volumetric (3D) data. In this paper we briefly provide background information and list some applications of PS filters from the literature. We then provide instructions and examples of how to use the classes and executables.