An Insight Toolkit (ITK) algorithm for adaptively thresholding images is presented in this paper. Currently, the usage of thresholding methods in ITK has made use of global thresholds, confidence connected thresholds and neighborhood strategies. The current work extends these family of filters by setting thresholds adaptively in local image regions. The user is not required to specify seed regions apriori which greatly eases the task of automatic segmentation. The thresholds are determined using Otsu's minimization of between-class variances in local image regions that are selected randomly throughout the domain. Using non-uniformly sampled thresholds, a continuous function is reconstructed throughout the image domain using a B-Spline approximation algorithm. Hence, the image domain is adaptively sampled by making use of the reconstructed threshold function. Most imaging modalities introduce some intensity inhomogeneities that can be recovered by this method. 

An Adaptive Thresholding Image Filter

This paper presents the results of the Head and Neck Auto-segmentationChallenge, which was part of the workshop “3D Segmentation in the Clinic: A Grand Challenge”. This workshop took place in London, UK, in September 2009, in conjunction with the 12th conference on Medical Image Computing and Computer Assisted Interventions (MICCAI). The aim of the challenge was to evaluate the performance of fully automated algorithms in segmenting the mandible and the brainstem in head and neck CT image data used in radiotherapy planning. We describe the motivation behind the clinical application selected for the challenge, the image data used, and the metrics applied for the quantitative assessment of the segmentation accuracy with respect to the ground truth segmentations provided by a clinical expert. The quantitative evaluation results of the auto-segmentations submitted by the workshop participants are included.

Head and Neck Auto-segmentation Challenge

This document introduces a program based on the algorithm described by Maria Cristina Tugurlan. The program uses file readers, image filters, and file writers from the Insight Toolkit ITK www.itk.org. It produces as output an image whose values are the times of first arrival of a wavefront that spreads from seed points with a speed at every point equal to the input image intensity. It performs the computation in parallel on distributed memory computers using the Message Passing Interface MPI. Each MPI process reads a small piece of the input image into memory. It computes fast marching on its piece. It sends and receives the values from fast marching at piece boundaries. It recalculates fast marching a number of times set from the command line using the new boundary values each time. Each MPI process writes only a small piece of the output file.

A substantial difference is seen when the output of MPI fast marching is compared to the output of serial fast marching. This difference may be acceptable for some uses. The program should be able to handle input images that are too large to fit in the memory of a single computer.

An Implementation of Parallel Fast Marching Using the Message Passing Interface

In this submission, we offer the GaussianInterpolationImageFunction which adds to the growing collection of existing interpolation algorithms in ITK for resampling scalar images such as the LinearInterpolateImageFunction, BSplineInterpolateImageFunction, and WindowedSincInterpolateImageFunction. We provide a brief discussion of the theory behind the submission and the algorithmic implementation.

Gaussian Interpolation

This guide is intended as a very brief introduction of the main tools in the Computational Morphometry Toolkit (CMTK), which is available in source code and as precompiled binaries from http://www.nitrc.org/projects/cmtk/. The target audience of this document are CMTK users, who might use this document as a reference to the most common processing tasks, and prospective users, who may find this information useful to determine whether CMTK provides functionality that they can use. We focus in particular on a simplified workflow for deformation morphometry studies based on magnetic resonance images: DICOM conversion, artifact correction, affine and nonlinear image registration, reformatting, Jacobian determinant map generation, and statistical hypothesis testing.

User Guide to The Computational Morphometry Toolkit

Although the KappaStatisticImageToImageMetric can be used to obtain the Dice metric (or mean overlap), there are other related measurements that are useful for evaluating results derived from various image analysis tasks. These measures include the target overlap, union overlap (or Jaccard coefficient), and false positive/negative errors. There are also related measures for multilabeled images. The class included with this submission, the itkLabelOverlapMeasuresImageFilter, is meant to provide an easy mechanism for calculating such measures.

Introducing Dice, Jaccard, and Other Label Overlap Measures To ITK

This document presents a set of classes (vtkPointSetOutlierRemoval, vtkPointSetNormalEstimation, vtkPointSetNormalOrientation, vtkPointSetCurvatureEstimation, vtkEuclideanMinimumSpanningTree, and vtkRiemannianGraphFilter) to enable several basic operations on point sets. These classes are implemented as VTK filters. Paraview plugin interfaces to the filters are also provided to allow extremely easy experimentation with the new functionality. We propose these classes as an addition to the Visualization Toolkit. 

The latest source code can be found here: https://github.com/daviddoria/PointSetProcessing

Point Set Processing for VTK - Outlier Removal, Curvature Estimation, Normal Estimation, Normal Orientation

RANdom SAmple Concensus (RANSAC) is an iterative method to estimate parameters of a model. It assumes that their are inliers in the data which are well explained by the chosen model. We propose a new class for VTK, vtkRANSACPlane, which estimates the best plane in a point set using the RANSAC method.

RANSAC Plane Fitting for VTK

We present a VTK-based implementation of our non-parametric vessel detection method that identifies vascular structures using a polar neighborhood profile. To accelerate the computationally intensive parts of the algorithm, we leverage the hardware capabilities in commodity graphics hardware using Compute Unified Device Architecture (CUDA). We present the results of our performance analysis and provide source code and examples to validate the reproducibility of our results.

A VTK-based, CUDA-optimized Non-Parametric Vessel Detection Method

The Wavefront OBJ format is a common file format for storing mesh data. It is convenient to be able to write VTK data structures out into this format, as the data can then be used in software that can read OBJ files.