Performing real-time interactive fiber tracking

Fast Automatic Segmentation of White Matter Streamlines Based on a Multi-Subject Bundle Atlas

This paper presents an algorithm for fast segmentation of white matter bundles from massive dMRI tractography datasets using a multisubject atlas. We use a distance metric to compare streamlines in a subject dataset to labeled centroids in the atlas, and label them using a per-bundle configurable threshold. In order to reduce segmentation time, the algorithm first preprocesses the data using a simplified distance metric to rapidly discard candidate streamlines in multiple stages, while guaranteeing that no false negatives are produced. The smaller set of remaining streamlines is then segmented using the original metric, thus eliminating any false positives from the preprocessing stage. As a result, a single-thread implementation of the algorithm can segment a dataset of almost 9 million streamlines in less than 6 minutes. Moreover, parallel versions of our algorithm for multicore processors and graphics processing units further reduce the segmentation time to less than 22 seconds and to 5 seconds, respectively. This performance enables the use of the algorithm in truly interactive applications for visualization, analysis, and segmentation of large white matter tractography datasets.

Real-time multi-peak tractography for instantaneous connectivity display

The computerized process of reconstructing white matter tracts from diffusion MRI (dMRI) data is often referred to as tractography. Tractography is nowadays central in structural connectivity since it is the only non-invasive technique to obtain information about brain wiring. Most publicly available tractography techniques and most studies are based on a fixed set of tractography parameters. However, the scale and curvature of fiber bundles can vary from region to region in the brain. Therefore, depending on the area of interest or subject (e.g., healthy control vs. tumor patient), optimal tracking parameters can be dramatically different. As a result, a slight change in tracking parameters may return different connectivity profiles and complicate the interpretation of the results. Having access to tractography parameters can thus be advantageous, as it will help in better isolating those which are sensitive to certain streamline features and potentially converge on optimal settings which are area-specific. In this work, we propose a real-time fiber tracking (RTT) tool which can instantaneously compute and display streamlines. To achieve such real-time performance, we propose a novel evolution equation based on the upsampled principal directions, also called peaks, extracted at each voxel of the dMRI dataset. The technique runs on a single Computer Processing Unit (CPU) without the need for Graphical Unit Processing (GPU) programming. We qualitatively illustrate and quantitatively evaluate our novel multi-peak RTT technique on phantom and human datasets in comparison with the state of the art offline tractography from MRtrix, which is robust to fiber crossings. Finally, we show how our RTT tool facilitates neurosurgical planning and allows one to find fibers that infiltrate tumor areas, otherwise missing when using the standard default tracking parameters.

Medical image processing on the GPU - past, present and future

Graphics processing units (GPUs) are used today in a wide range of applications, mainly because they can dramatically accelerate parallel computing, are affordable and energy efficient. In the field of medical imaging, GPUs are in some cases crucial for enabling practical use of computationally demanding algorithms. This review presents the past and present work on GPU accelerated medical image processing, and is meant to serve as an overview and introduction to existing GPU implementations. The review covers GPU acceleration of basic image processing operations (filtering, interpolation, histogram estimation and distance transforms), the most commonly used algorithms in medical imaging (image registration, image segmentation and image denoising) and algorithms that are specific to individual modalities (CT, PET, SPECT, MRI, fMRI, DTI, ultrasound, optical imaging and microscopy). The review ends by highlighting some future possibilities and challenges.

Accelerating fibre orientation estimation from diffusion weighted magnetic resonance imaging using GPUs

With the performance of central processing units (CPUs) having effectively reached a limit, parallel processing offers an alternative for applications with high computational demands. Modern graphics processing units (GPUs) are massively parallel processors that can execute simultaneously thousands of light-weight processes. In this study, we propose and implement a parallel GPU-based design of a popular method that is used for the analysis of brain magnetic resonance imaging (MRI). More specifically, we are concerned with a model-based approach for extracting tissue structural information from diffusion-weighted (DW) MRI data. DW-MRI offers, through tractography approaches, the only way to study brain structural connectivity, non-invasively and in-vivo. We parallelise the Bayesian inference framework for the ball & stick model, as it is implemented in the tractography toolbox of the popular FSL software package (University of Oxford). For our implementation, we utilise the Compute Unified Device Architecture (CUDA) programming model. We show that the parameter estimation, performed through Markov Chain Monte Carlo (MCMC), is accelerated by at least two orders of magnitude, when comparing a single GPU with the respective sequential single-core CPU version. We also illustrate similar speed-up factors (up to 120x) when comparing a multi-GPU with a multi-CPU implementation.

GPU-based acceleration of an automatic white matter segmentation algorithm using CUDA

This paper presents a parallel implementation of an algorithm for automatic segmentation of white matter fibers from tractography data. We execute the algorithm in parallel using a high-end video card with a Graphics Processing Unit (GPU) as a computation accelerator, using the CUDA language. By exploiting the parallelism and the properties of the memory hierarchy available on the GPU, we obtain a speedup in execution time of 33.6 with respect to an optimized sequential version of the algorithm written in C, and of 240 with respect to the original Python/C++ implementation. The execution time is reduced from more than two hours to only 35 seconds for a subject dataset of 800,000 fibers, thus enabling applications that use interactive segmentation and visualization of small to medium-sized tractography datasets.

CUDA-Accelerated Geodesic Ray-Tracing for Fiber Tracking

Diffusion Tensor Imaging (DTI) allows to noninvasively measure the diffusion of water in fibrous tissue. By reconstructing the fibers from DTI data using a fiber-tracking algorithm, we can deduce the structure of the tissue. In this paper, we outline an approach to accelerating such a fiber-tracking algorithm using a Graphics Processing Unit (GPU). This algorithm, which is based on the calculation of geodesics, has shown promising results for both synthetic and real data, but is limited in its applicability by its high computational requirements. We present a solution which uses the parallelism offered by modern GPUs, in combination with the CUDA platform by NVIDIA, to significantly reduce the execution time of the fiber-tracking algorithm. Compared to a multithreaded CPU implementation of the same algorithm, our GPU mapping achieves a speedup factor of up to 40 times.