An Evolutionary Approach for Automatic Seedpoint Setting in Brain Fiber Tracking

In this paper we present an evolutionary approach for optimising the seedpoint setting in brain fiber tracking. Our aim is to use Diffusion Tensor Imaging (DTI) data and Diffusion Magnetic Resonance Imaging (dMRI) data for feeding an automatic fiber tracking approach. Our work focusses on customising an evolutionary algorithm to find nerve fibers within diffusion data and allocate an appropriate number of seedpoints to them. This is necessary for the subsequent fiber reconstruction algorithms to work. The algorithm considerably enhances the speed and quality of the reconstruction and proves to be promising in leading to an automatic fiber tracking procedure used in medical imaging.