Multi-scale filters implemented by cellular automaton for retinal layers modelling

In the following of our works, this article presents the use of cellular automata, as part of retinal layer modelling. The main innovation of this work is the computing method based on a new approach using multi-scale filtering without loss of resolution. This new approach provides us a solution to the full-resolution filtering method that was previously too expensive in computing time. More precisely the implementation of the retinal horizontal layer penalised us in our modelling of the whole retinal architecture. The efficiency of this multi-scale model is demonstrated, and the treatment is described and applied to static and dynamic stimuli of real scenes.