Cellular Associative Symbolic Processing for Pattern Recognition

A cellular network of associative processors capable of symbolic processing is described in this paper. The processors are placed in a two dimensional array and they can perform a global set of symbolic rules deening their next state and the messages to be passed to their neighbours. The system follows a cellular automata like operation and the aim is to transform initial symbolic descriptions of patterns to the corresponding object level ones in order to identify patterns in complex or noisy scenes. A learning algorithm based on a hierarchical approach for the description of the structure of the patterns is used to produce the symbolic rules while the underlying symbolic processing engine of the system is the AURA type of associative memory. The use of this model enables the system to operate in high speed and also allows direct hardware implementation.