The following participants submitted the results for SSGCI competition before the deadline:
Product Graph-based Higher Order Contextual Similarities for Inexact Subgraph Matching
Anjan Dutta(1), Josep Lladós(1), Horst Bunke(2) and Umapada Pal(3)
(1) Computer Vision Center, Universitat Autònoma de Barcelona, Barcelona, Spain
(2) Institute of Computer Science and Applied Mathematics, University of Bern, Bern, Switzerland
(3) Computer Vision and Pattern Recognition Unit, Indian Statistical Institute, Kolkata, India Email:
We propose contextual similarities between a pair of nodes each coming from two graphs to be matched. This is done by considering their tensor product graph (TPG), where each node is an ordered pair of nodes of the operand graphs. Contextual similarities between a pair of nodes are computed by accumulating weighted walks terminating at their paired node in TPG. Once the contextual similarities are obtained, we formulate subgraph matching as a node and edge selection problem in TPG. We use contextual similarities to construct an objective function and optimize it with a linear programming approach.
Description of the method submitted to the Competition on Subgraph Spotting in Graph representation of Comic Book Images
Pierre Héroux and Sébastien Adam
LITIS, University of Rouen, France
The proposed method relies on the computation of the minimum cost subgraph matching.
This problem is modeled as a binary linear program were variables represent edit operations. Each edit operation is penalized with cost function. A setting of this cost functions defines the relative weights of each label attribute in the penalty computation.
In order to return a result in a reasonable time, a non optimal vertex association is returned, which corresponds to the best solution after a partial exploration of the solution tree.
The submitted results are a combination of results obtained with four different settings of the cost functions.