Image compression using stochastic neural networks

Image compression using stochastic artificial neural networks (SANNs) is studied. The ideal is to store an image in a stable distribution of a stochastic neural network. Given an input image f εF, one can find a SANN t ε T such that the equilibrium distribution of this SANN is the given image f. Therefore, the input image, f, is encoded into a specification of a SANN, t. This mapping from F (image space) to T (parameter space of SANN) defines the SANN transformation. It is shown that the compression ratio R of the SANN transformation is R=O(n/(K (log n)²)) where n is the number of pixels. To complete a SANN transformation, SANN equations must be solved. Two SANN equations are presented. The solution of SANN is briefly discussed