A GPU-based implementation of the MRF algorithm in ITK package

The analysis of medical image, in particular Magnetic Resonance Imaging (MRI), is a very useful tool to help the neurologists on the diagnosis. One of the stages on the analysis of MRI is given by a classification based on the Markov Random Fields (MRF) method. It is possible to find in the literature several packages to carry out this analysis, and of course, the classification tasks. One of them is the Insight Segmentation and Registration Toolkit (ITK). The analysis of MRI is an expensive computational task. In order to reduce the execution time spent on the analysis of MRI, parallelism techniques can be used. Currently, Graphics Processing Units (GPUs) are becoming a good choice to reduce the execution time of several applications at a low cost. In this paper, the authors present a GPU-based classification using MRF from the sequential implementation that appears in the ITK package. The experimental results show a spectacular execution time reduction being the GPU-based implementation up to 118 times faster than the sequential implementation included in the ITK package. Moreover, this result is also observed by reducing the total power consumption in a significant amount.