A new model-based spherical deconvolution method for multi-fiber reconstruction

We propose a novel method to reconstruct the multi-fiber orientation in this work. Firstly, we convolve DW-MRI signal with response function which is model-based to get the discretized fiber orientation distribution function. Secondly, we used the centered spherical Gaussian functions and the discretized fiber orientation density function to form a continuous fiber orientation density function which can correctly reveal the true fiber orientation density distributions in the 3D space. Finally, the properties of the method are demonstrated using both simulations and real datasets. And the results demonstrated the superiority of the proposed method over several existing multi-fiber reconstruction methods.