Second-Level Upperbound Screening
Use of a more accurate upperbound calculation that uses an assignment procedure to yield a tighter bound on |E(G12)|
Based on mapping each node in G1 to each possible matching node in G2 , and counting the number of possible matching edges that would result from each such mapping
The linear assignment identifies a (not necessarily unique) set of mappings of individual nodes from G1 to individual nodes from G2 that maximises the number of possible matching edges