mexSphericalPolarFourierImaging
===============================

.. code-block:: none

  % mexSphericalPolarFourierImaging: estimate continuous DWI signal and EAP represented by coefficients of given basis. 
  %
  % Usage:
  %   [signal] = mexSphericalPolarFourierImaging(dwi4DNoramlized, grad, bVec, params)
  % 
  % Inputs:
  %   dwi4DNoramlized:       4D DWI data, normalized by b0 image;
  %   grad :                 orientation matrix where each row is an orientation (x,y,z), caresian format
  %   bVector:               b values in a vector, each row is a b value
  %
  %   params.MD0 :           Typical mean diffusivity (MD) value. The default value is 0.7e-3. 
  %   params.tau :           Tau value. The default is calculated based on 4*pi*pi*tau=1.
  %   param.scale :          Scale value for the SPF basis in radial part, the default value is set by tau and MD0 and tau 
  %   params.sh:             SH order
  %   params.ra:             order in radial part
  %   params.lambdaSH :      lambda in spherical part
  %   params.lambdaRA :      lambda in radial part
  %   params.lambdaL1 :      lambda for L1 solver
  %   params.odfOrder :      odf order 
  %   params.radius   :      Radius for EAP profile.
  %   params.mdImage  :      Mean diffusivity Image for adaptive scale.
  %   params.estimation :    (LS, L1_2, L1_DL) LS: least square. 
  %                          L1_2 means Laplace-Beltrami weight (two regularization parameters lambdaSH and lambdaRA); 
  %                          L1_DL only use lambdaL1, because the learned dictionary corresponses the regularization matrix.
  %   params.solver   :      (FISTA_LS, SPAMS)
  %                          FISTA_LS: FISTA using least square initialization.  
  %                          SPAMS (default): use spams' weighted lasso solver. 
  %   params.maxIter  :      Maximal number of iteration in L1 FISTA.
  %   params.minChange :     Minimal change percentage of the cost function and variable for l1 oprimization.
  %   params.mask :          Mask file. 2D or 3D matrix. 
  %
  %   params.thread :        Number of thread. Default is -1, which means it is automatically determined.
  %   params.verbose :       Verbose level. 0: no output log. 1: normal log. 2: large log for debug.
  %
  % Outputs:
  %   signal:                coefficient under a given basis, 4D image
  %
  %
  % Reference: 
  %  1. "Model-Free, Regularized, Fast, and Robust Analytical Orientation Distribution Function Estimation",
  %     Jian Cheng, Aurobrata Ghosh, Rachid Deriche, Tianzi Jiang, Medical Image Computing and Computer-Assisted Intervention (MICCAI'10), vol. 6361, pp. 648–656, sep, 2010.
  %  2. "Model-free and Analytical EAP Reconstruction via Spherical Polar Fourier Diffusion MRI", 
  %     Jian Cheng, Aurobrata Ghosh, Tianzi Jiang, Rachid Deriche, Medical Image Computing and Computer-Assisted Intervention (MICCAI'10), vol. 6361, pp. 590–597, sep, 2010. 
  %
  % Copyright (c) 2013, Jian Cheng <jian.cheng.1983@gmail.com>
  %