ABSTRACT
We present a novel self-organizing structure recognition (SOSR) network for classification and recognition of relational structures represented by graphs. The system consists of several subnets each comparing an input structure with a given model structure. The subnets are indirectly coupled via a winner-take-all (WTA) classifier. During classification the SOSR system deactivates subnets which indicate large dissimilarities between the input structure and the corresponding models. First experiments show that this mechanism significantly reduces the computational effort in comparison to traditional classification systems using a comparative maximum selector as a classifier.
