Bipartite graph labeling for the subcircuit recognition problem

Subcircuit recognition (SR) is a problem of recognition of a small model bipartite graph (BG) associated with a subcircuit in a larger object BG associated with a circuit. The performance of the SR methods strongly depends on labeling both the BGs. Unfortunately, known labeling algorithms possess weak discriminative abilities since in forming labels they exploit the structure of local vertex connections mainly. As a result, we have to use computationally expensive search-based algorithms for SR. In this paper we propose a new efficient algorithm to label BG vertices for the SR problem. This algorithm is based on iterative hashing vertex labels according to the special rules. The new method takes into account the structure of non-local vertex surroundings and, as a result, it possesses good discriminative abilities. This allows combining it with an efficient probabilistic graph recognition algorithm to solve the SR problem. The experiments show that using the new labeling algorithm substantially improves the SR results in comparison with standard labeling methods