2015 | OriginalPaper | Buchkapitel
Attributed Relational Graph Matching with Sparse Relaxation and Bistochastic Normalization
verfasst von : Bo Jiang, Jin Tang, Bin Luo
Erschienen in: Graph-Based Representations in Pattern Recognition
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
Attributed relational graph (ARG) matching problem can usually be formulated as an Integer Quadratic Programming (IQP) problem. Since it is NP-hard, relaxation methods are required. In this paper, we propose a new relaxation method, called Bistochastic Preserving Sparse Relaxation Matching (BPSRM), for ARG matching problem. The main benefit of BPSRM is that the mapping constraints involving both discrete and bistochastic constraint can be well incorporated in BPSRM optimization. Thus, it can generate an approximate binary solution with one-to-one mapping constraint for ARG matching problem. Experimental results show the effectiveness of the proposed method.