how to train the deep neural network with GLCM features to find out whether an input image is stego or clean ?(based on these features)

Hi,

i am trying to do a project where i have to train the deep neural network with glcm features and based on this glcm features , the DNN should be able to identify whether an image is clean or data hidden one .

i have extracted four features using glcm . the values for each image are like

contrast =[5.81944444444445,6.47777777777778] homogenity=[0.737242345924548,0.706970684878285] correlation=[0.279660493827161,0.274506172839506] energy=[0.799867724867725,0.798429232804233]

i have made an algorithm based on the deep learn toolbox , but dont know where to implement these features in it.i have uploaded the algorithm file in this .

please anyone have a look and help me with it ?