Improvement of DT-MRI Processing Algorithms using Neural Networks

This paper deals with the development of several neural network (NN)-based schemes that can be applied to medical image processing systems. Specifically, we are interested in the estimation of fiber bundles (fiber tracking) in diffusion tensor (DT) fields acquired via magnetic resonance imaging (MRI). In previous work, we proposed a tracking scheme that was successfully tested with synthetic and real DT-MRI images. In this paper, a NN-based scheme for tuning-up these tracking systems is introduced and tested. This issue has been traditionally undertaken under an heuristical approach. Besides, several novel simplification schemes are developed so as to reduce the computational complexity of the Bayesian approaches for fiber tracking. The proposed procedures have been numerically evaluated. The efficient tracking of white matter fibers in the human brain will improve the diagnosis and treatment of many neuronal diseases.