Christophe Irniger

Distance measures for image segmentation evaluation

The task considered in this paper is performance evaluation of region segmentation algorithms in the ground-truth-based paradigm. Given a machine segmentation and a ground-truth segmentation, performance measures are needed. We propose to consider the image segmentation problem as one of data clustering and, as a consequence, to use measures for comparing clusterings developed in statistics and machine learning. By doing so, we obtain a variety of performance measures which have not been used before in image processing. In particular, some of these measures have the highly desired property of being a metric. Experimental results are reported on both synthetic and real data to validate the measures and compare them with others.

Graph matching – challenges and potential solutions

Structural pattern representations, especially graphs, have advantages over feature vectors. However, they also suffer from a number of disadvantages, for example, their high computational complexity. Moreover, we observe that in the field of statistical pattern recognition a number of powerful concepts emerged recently that have no equivalent counterpart in the domain of structural pattern recognition yet. Examples include multiple classifier systems and kernel methods. In this paper, we survey a number of recent developments that may be suitable to overcome some of the current limitations of graph based representations in pattern recognition.

Theoretical analysis and experimental comparison of graph matching algorithms for database filtering

In structural pattern recognition, an unknown pattern is often transformed into a graph that is matched against a database in order to find the most similar prototype in the database. Graph matching is a powerful yet computationally expensive procedure. If the sample graph is matched against a large database of model graphs, the size of the database is introduced as an additional factor into the overall complexity of the matching process. Database filtering procedures are used to reduce the impact of this additional factor. In this paper we report the results of a basic study on the relation between filtering efficiency and graph matching algorithm performance, using different graph matching algorithms for isomorphism and subgraph-isomorphism.

Computer network monitoring and abnormal event detection using graph matching and multidimensional scaling

Computer network monitoring and abnormal event detection have become important areas of research. In previous work, it has been proposed to represent a computer network as a time series of graphs and to compute the difference, or distance, of consecutive graphs in such a time series. Whenever the distance of two graphs exceeds a given threshold, an abnormal event is reported. In the present paper we go one step further and compute graph distances between all pairs of graphs in a time series. Given these distances, a multidimensional scaling procedure is applied that maps each graph onto a point in the two-dimensional real plane, such that the distances between the graphs are reflected, as closely as possible, in the distances between the points in the two-dimensional plane. In this way the behaviour of a network can be visualised and abnormal events as well as states or clusters of states of the network can be graphically represented. We demonstrate the feasibility of the proposed method by means of synthetically generated graph sequences and data from real computer networks.

Graph database filtering using decision trees

Graphs are a powerful representation formalism for structural data. They are, however, very expensive from the computational point of view. In pattern recognition it is often necessary to match an unknown sample against a database of candidate patterns. In this process, however, the size of the database is introduced as an additional factor into the overall complexity of the matching process. To reduce the influence of that factor, an approach based on machine learning techniques is proposed in this paper. Graphs are represented using feature vectors. Based on these vectors a decision tree is built to index the database. The decision tree allows at runtime to eliminate a number of graphs from the database as possible matching candidates.

Using attributed plex grammars for the generation of image and graph databases

In this paper, a methodology for the generation of benchmarks in pattern recognition is described. The patterns are represented by means of an attributed plex language, which are based on plex grammars augmented by attributes. It is shown that the generated patterns are particularly suitable for the extraction of graph-based representations. As a result, databases of artificial pictures and correspondent graphs can be generated. These collections of graphs are very appropriate for benchmarks in the area of structural pattern recognition, since they are originated from a grammar and not from random distributions. The tools for creating the databases are public domain and have been already used for benchmarking artificial neural networks operating on structured domains.

Image segmentation evaluation by techniques of comparing clusterings

The task considered in this paper is performance evaluation of region segmentation algorithms in the ground truth (GT) based paradigm. Given a machine segmentation and a GT reference, performance measures are needed. We propose to consider the image segmentation problem as one of data clustering and, as a consequence, to use measures for comparing clusterings developed in statistics and machine learning. By doing so, we obtain a variety of performance measures which have not been used before in computer vision. In particular, some of these measures have the highly desired property of being a metric. Experimental results are reported on both synthetic and real data to validate the measures and compare them with others.

Recovery of missing information in graph sequences by means of reference pattern matching and decision tree learning

Algorithms for the analysis of graph sequences are proposed in this paper. In particular, we study the problem of recovering missing information and predicting the occurrence of nodes and edges in time series of graphs. Two different recovery schemes are developed. The first scheme uses reference patterns that are extracted from a training set of graph sequences, while the second method is based on decision tree induction. Our work is motivated by applications in computer network analysis. However, the proposed recovery and prediction schemes are generic and can be applied in other domains as well.

Decision Tree Structures for Graph Database Filtering

In structural pattern recognition it is often required to match an unknown sample against a database of candidate patterns in order to find the most similar prototype. If the patterns are represented using graphs, the sample’s graph is matched against a database of model graphs and the pattern recognition problem is turned into a graph matching problem. Graph matching is a powerful yet computationally expensive procedure. If the unknown sample is matched against a whole database of prototypes, the size of the database is introduced as an additional factor into the overall complexity of the matching process. To reduce the influence of that factor an approach based on machine learning techniques is proposed in this paper. The graphs are represented using feature vectors. Based on these vectors a decision tree is built to index the database. The decision tree allows at runtime to eliminate a number of graphs from the database as possible matching candidates. Experimental results are reported demonstrating the efficiency of the proposed filtering procedure. The work presented in this paper extends previous studies from the case of graph-isomorphism to the problem of subgraph-isomorphism.

Decision trees for error-tolerant graph database filtering

An important topic in pattern recognition is retrieval of candidate patterns from a database according to a given sample input pattern. Using graphs, the database retrieval problem is turned into a graph matching problem. In this paper we propose a method based on decision trees to filter a database of graphs according to a given input graph. The present paper extends previous work concerned with graph and subgraph isomorphism to the case of error-tolerant graph matching.