This work is integrated in the MICCAI Grand Challenge: MR Brain Image Segmentation 2013. It aims for the automatic segmentation of brain into Cerebrospinal fluid (CSF), Gray matter (GM) and White matter (WM). The provided dataset contains patients with white matter lesions, which makes the segmentation task more challenging. The proposed algorithm uses multi-sequence MR images to extract meaningful features and learn a Random Decision Forest that classifies each voxel of the image. The results show that it is robust to the presence of the white matter lesions, and the metrics show that the overall results are competitive.

Automatic Brain Tissue Segmentation of Multi-sequence MR Images Using Random Decision Forests

An automated method of MR Brain image segmentation is presented. A block based Expectation Maximization method is presented for the tissue classification of MR Brain images. The standard Gaussian Mixture Model is the most widely used method for MR Brain Image Segmentation and Expectation Maximization algorithm is used to estimate the model parameters. The Gaussian Mixture Model considers each pixel as independent and does not take into account the spatial correlation between the neighbouring pixels. Hence the segmentation result obtained using standard GMM is highly sensitive to Inensity Non-Uniformity and noise. The image is divided into blocks before applying EM since the GMM is preserved in the local image blocks. Also, Nonsubsampled Contourlet Transform is employed to incorporate the spatial correlation among the neighbouring pixels. The method is applied to the 12 MR Brain volumes of MRBRAINS13 test data and the White Matter, Gray Matter and CSF structures were segmented.

Modified Expectation Maximization Method for Automatic Segmentation of MR Brain Images

In this paper, we validate our proposed segmentation algorithm called Bayesian-based adaptive mean-shift (BAMS) on real mul-timodal MR images provided by the MRBrainS challenge. BAMS is a fully automatic unsupervised segmentation algorithm. It is based on the adaptive mean shift wherein the adaptive bandwidth of the kernel for each feature point is estimated using our proposed Bayesian approach [1]. BAMS is designed to segment the brain into three tissues; white matter (WM), gray matter (GM) and cerebrospinal fluid (CSF). The performance of the algorithm is evaluated relative to the manual segmentation (ground truth). The results of our proposed algorithm show the average Dice index 0.8377±0.036 for the WM, 0.7637±0.038 for the GM and 0.6835 ±0.023 for the CSF.

Multimodal MR Brain Segmentation Using Bayesian-based Adaptive Mean-Shift (BAMS)

Characterization of anatomical structure of the brain and efficient algorithms for automatically analyzing brain MRI have gained an increasing interest in recent years. In this paper, we propose an algorithm that automatically segments the anatomical structures of magnetic resonance human brain images. Our method uses the prior knowledge of labels given by experts to statistically investigate the spatial correspondences of brain structures in subject images. We create a multi-atlas by registering the training images to the subject image and then propagating corresponding labels to the graph of the image. Label fusion then combines these multiple labels of atlases into one label at each voxel with intensity similarity based weighted voting. Finally we cluster the graph using multiway cut in order to achieve the final 3D segmentation of the subject image. The promising evaluation results of our segmentation method on the MRBrainS13 Test Dataset shows the efficiency and robustness of our algorithm.

Multi-Atlas Brain MRI Segmentation with Multiway Cut

We present a new filter for the Insight Toolkit (ITK) for reducing the resolution of an image by an integer factor while averaging called BinShrink. This filter provides a new level of performance to ITK for reducing resolution and noise present in an image. The filter supports streaming, multi-threading and most of ITK's pixel types including scalars, Vectors, SymmetricSecondRankTensors, and RGBPixels. The filter has been optimized to efficiently access the input image thereby greatly increasing performance over conventional methods.

BinShrink: A multi-resolution filter with cache efficient averaging.

This document describes the implementation of an ITK class to support the reading and writing of Meshes in STL file format. The Meshes are assumed to contain 2D manifolds embedded in a 3D space. In practice, it would be desirable to use this class mostly to read and write QuadEdgeMeshes.

This paper is accompanied with the source code, input data, parameters and output data that the authors 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.

STL file format MeshIO class for ITK

This article describes implementation of a CLI module to merge an image into the space of another adjacent image using the Insight Toolkit. The algorithm constructs a smoothly stitched image out of two or more smaller images. The overlapping regions of the images are used to align the images. The registration method is derived from ITK's Registration framework. The algorithm works with 2D and 3D images.

Using TubeTK to Automatically Merge Volumes

Several applications such as multi-atlas segmentation require frequent iteration over multiple image volumes at the same time.
Doing so with the regular ITK iterators is tedious and error prone as it requires updating each iterator at end of each iteration. Failing to do so results in hard to debug errors and crashes.
The MultipleImageIterator is a simple wrapper class that tries to make this more convenient.


A MultipleImageIterator for iterating over multiple images simultaneously

Three-dimensional laser scanner is a modern technology used to detect and capture the shape of complex objects. There are a lot of possible application fields for this technology: applications in the industrial sector (reverse engineering), in the medical field or in the cultural heritage field for virtual reconstruction of works of art. 
The main goal of this work is to make accessible the three-dimensional models obtained from laser scans on mobile devices (smartphones, tablets), exploiting a remote rendering application on high-performance workstations.
This paper also shows how this can be done fairly easily using one of the most popular instruments in the field of scientific visualization, VTK.

VTK Remote Rendering of 3D Laser Scanner Ply files for Android Mobile Devices

In this article, we present an implementation of a flexible framework for non-parametric variational image registration, realized as part of ITK's finite difference solver hierarchy.

In a variational registration setting, the transformation is found by minimizing an energy functional that consists of (at least) two terms: a distance measure between fixed and transformed moving image and a smoothness condition for the transformation. The specific choice of these terms depends on the particular application and requires consideration of, for example, image content and imaging modalities. Following this view, the presented framework can be seen as a generalization of the demons algorithm in which two key aspects remain exchangeable: the force term (or registration function, according to the distance measure) and the regularizer (according to the smoothness condition). Moreover, two transformation models are realized: either a dense displacement field or a stationary velocity field to restrain the transformation to the space of diffeomorphisms.

In its current state, the framework includes implementations of forces based on the Sum of Squared Differences (SSD), Normalized Cross Correlation (NCC) and the demons algorithm as well as Gaussian, diffusion and elastic smoothing. However, the implementation of further components is possible and encouraged.
