Deterministic search for relational graph matching

摘要

This paper describes a comparative study of various deterministic discrete search-strategies for graph-matching. The framework for our study is provided by the Bayesian consistency measure recently reported by Wilson and Hancock (IEEE PAMI 19 (1997) 634-648; Pattern Recognition 17 (1996) 263-276) and Wilson et al. (Comput. Vision Image Understanding 72 (1998) 20-38') We investigate two classes of update process. The first of these aims to exploit discrete gradient ascent methods. We investigate the effect of searching in the direction of both the local and global gradient maximum. An experimental study demonstrates that although more computationally intensive, the global gradient method offers significant performance advantages in terms of accuracy of match. Our second search strategy is based on tabu search. In order to develop this method we introduce memory into the search procedure by defining context-dependant search paths. We illustrate that although it is more efficient than the global gradient method, tabu search delivers almost comparable performance.