Sébastien Sorlin

Reactive tabu search for measuring graph similarity

Graph matching is often used for image recognition. Different kinds of graph matchings have been proposed such as (sub)graph isomorphism or error-tolerant graph matching, giving rise to different graph similarity measures. A first goal of this paper is to show that these different measures can be viewed as special cases of a generic similarity measure introduced in [8]. This generic similarity measure is based on a non-bijective graph matching (like [4] and [2]) so that it is well suited to image recognition. In particular, over/under-segmentation problems can be handled by linking one vertex to a set of vertices. In a second part, we address the problem of computing this measure and we describe two algorithms: a greedy algorithm, that quickly computes sub-optimal solutions, and a reactive Tabu search algorithm, that may improve these solutions. Some experimental results are given.

A parametric filtering algorithm for the graph isomorphism problem

We introduce a new filtering algorithm, called IDL(d)-filtering, for a global constraint dedicated to the graph isomorphism problem—the goal of which is to decide if two given graphs have an identical structure. The basic idea of IDL(d)-filtering is to label every vertex with respect to its relationships with other vertices around it in the graph, and to use these labels to filter domains by removing values that have different labels. IDL(d)-filtering is parameterized by a positive integer value d which gives a limit on the distance between a vertex to be labelled and the set of vertices considered to build its label. We experimentally compare different instantiations of IDL(d)-filtering with state-of-the-art dedicated algorithms and show that IDL(d)-filtering is more efficient on regular sparse graphs and competitive on other kinds of graphs.

Filtering for subgraph isomorphism

A subgraph isomorphism problem consists in deciding if there exists a copy of a pattern graph in a target graph. We introduce in this paper a filtering algorithm dedicated to this problem. The main idea is to label every node with respect to its relationships with other nodes of the graph, and to define a partial order on these labels in order to express compatibility of labels for subgraph isomorphism. This partial order over labels is used to filter domains. Labelings can also be strengthened by adding information from the labels of the neighbors. Such a strengthening can be applied iteratively until a fixpoint is reached. Practical experiments illustrate that our new filtering approach is more effective on difficult instances of scale free graphs than state-of-the-art algorithms and other CP approaches.

A Generic Graph Distance Measure Based on Multivalent Matchings

HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. A Generic Graph Distance Measure Based on Multivalent Matchings Sébastien Sorlin, Christine Solnon, Jean-Michel Jolion

A comparative study of ant colony optimization and reactive search for graph matching problems

Many applications involve matching two graphs in order to identify their common features and compute their similarity. In this paper, we address the problem of computing a graph similarity measure based on a multivalent graph matching and which is generic in the sense that other well known graph similarity measures can be viewed as special cases of it. We propose and compare two different kinds of algorithms: an Ant Colony Optimization based algorithm and a Reactive Search. We compare the efficiency of these two algorithms on two different kinds of difficult graph matching problems and we show that they obtain complementary results.