% Training data preparison by calling MRH_trainingPrep.
%After running this demo, the data and label file
% in .mat form like traindataJG_allMRIs_fluo.mat data label -v7.3; will be saved in
% current workspace.

%  - Zifei Liang (zifei.liang@nyumc.org)
% Using code please refer our work: 
% Inferring Maps of Cellular Structures from MRI Signals using Deep Learning 
% https://www.biorxiv.org/content/10.1101/2020.05.01.072561v1

clc;clear; close all;
% Please replace by the training data location by USER
work_folder = ['.\Train_Data\Subj\'];
halfpatch_size = 1;
% initial sample step over pixels. higher leads to fewer training samples. 
stride0 = 20;
% location of the target histology .img. Here we set one autofluorescence as target. 
% Target can be replaced by the imgs in foder
% .\Train_Data\autofluorescence\
fluo_img = load_untouch_nii(['.\Train_Data\Allen_Autofluo\AllenPathology2P60.img']);
% call training patches Prep function.
[data,label] = MRH_trainingPrep(work_folder,halfpatch_size,stride0, fluo_img);