==================
**  files:
==================


**************data***************       ********************************explanation**************************************

IMAGES.mat (512*512*10)                 : The benchmark database from Olshausen and Field (1996). In the database, ten
                                          natural images were preprocessed with a pseudo-whitening filter, which flattens 
                                          the spectrum of natural images by rescaling Fourier coefficients. 

train_patches.mat (100000*196)          : 100000 small patches (size 14×14) were extracted from random positions of the ten 
                                          whitened images. Furthermore, a sigmoid function was applied to the pixel intensities 
                                          to fit their values into the range [0, 1].

learned_DiverseRBM.mat (structure)      : the learned diverse RBM model 
                                          learned_DiverseRBM.W        ... The weights of the connections
                                          learned_DiverseRBM.b        ... The biases of the hidden layer
                                          learned_DiverseRBM.c        ... The biases of the visible layer
**********************************      **********************************************************************************



************* m.file *************      *******************************explanation****************************************

train_DiverseRBM.m                      : the main traiing algoirthm  
image_reconstruction.m                  : the code for the image reconstruction experiments 
display_patches_weight.m                : the code for displaying the weight W in a model 


RF_demo.m                               : visulaize the receptive fields (RFs) of a learned RBM;
img_reconst_demo.m                      : visualize the image reconstruction experiments 
**********************************      **********************************************************************************






====================
** run the code 
====================

>> RF_demo;              % visulaize the receptive fields (RFs) of a learned RBM;

>> img_reconst_demo;     % visualize the image reconstruction experiments 

