Outlier robust point correspondence based on GNCCP

• Point correspondence is formulated as the minimization of a quadratic objective.
• It directly targets at finding a specified number of best vertex assignments.
• GNCCP, a combinatorial optimization framework, is used to optimize the objective.
• The method can be applied to both undirected and directed graphs.

Graph matching is a fundamental problem in pattern recognition and computer vision. In this paper we introduce a novel graph matching algorithm to find the specified number of best vertex assignments between two labeled weighted graphs. The problem is first explicitly formulated as the minimization of a quadratic objective function and then solved by an optimization algorithm based on the recently proposed graduated nonconvexity and concavity procedure (GNCCP). Simulations on both synthetic data and real world images witness the effectiveness of the proposed method.