Figure 4
From: Reliable Steganalysis Using a Minimum Set of Samples and Features
Structure of a classical Single Layer Feedforward Neural Network (SLFN). The input values (the data) 
are weighted by the 
coefficients. A possible bias 
(not on the figure) can be added to the weighted inputs 
. An activation function 
taking this weighted inputs (plus bias) as input is finally weighted by output coefficients 
to obtain the output 
.