41
(iii) Direct template derivation method:
The third method is the direct template design for those desired functions that are exactly
explicit. This method is accurate but it is not always trivially possible to map any desired
function onto the CNN system model. Depending on the function, enhancing the CNN is
possible, such as adding a new layer or a specific nonlinear term [98]. We know that there
are many application based on different PDE model, such as inpainting for recovering
corrupted regions in image [70], image segmentation [71], noise reduction edge
preservation [72]. The procedure of solving PDEs in CNN is by transforming a PDE to set of
ODEs as a coupled system. After transforming a continuous spatial PDE to an array of
discrete interactive systems which are ODEs, we can map it on CNN cells. Because CNN is
natural and flexible paradigm for modeling a simple locally interconnected dynamical
system which are grid base. Detail of this template modeling is already described in sub-
chapter 3-3.
Our goal in this chapter is to give a practical introduction to template design. We however
focus on the heuristic method based on genetic algorithms.
