An Optimal Graph Theoretic Approach to Data Clustering: Theory and Its Application to Image Segmentation

  title={An Optimal Graph Theoretic Approach to Data Clustering: Theory and Its Application to Image Segmentation},
  author={Zhenyu Wu and Richard M. Leahy},
  journal={IEEE Trans. Pattern Anal. Mach. Intell.},
  • Zhenyu Wu, R. Leahy
  • Published 1 November 1993
  • Computer Science
  • IEEE Trans. Pattern Anal. Mach. Intell.
A novel graph theoretic approach for data clustering is presented and its application to the image segmentation problem is demonstrated. The data to be clustered are represented by an undirected adjacency graph G with arc capacities assigned to reflect the similarity between the linked vertices. Clustering is achieved by removing arcs of G to form mutually exclusive subgraphs such that the largest inter-subgraph maximum flow is minimized. For graphs of moderate size ( approximately 2000… 

Figures and Tables from this paper

Image Segmentation With Maximum Cuts
Different graph structures that can be used are defined and it is shown that the computational complexity of the problem significantly depends on the structure of the graph.
Image Segmentation Based on Minimal Spanning Tree and Cycles
  • T. N. Janakiraman, P. Mouli
  • Computer Science
    International Conference on Computational Intelligence and Multimedia Applications (ICCIMA 2007)
  • 2007
A novel graph theoretic approach for image segmentation that is able to locate almost proper region boundaries of clusters and is applicable to any image domain is presented.
Agglomerative connectivity constrained clustering for image segmentation
  • Jia Li
  • Computer Science
    Stat. Anal. Data Min.
  • 2011
A new fully automatic segmentation algorithm for color photographs that combines the top‐down k‐means clustering with the bottom‐up agglomerative clustering method and finds that it is advantageous to conduct clustering at multiple stages through which the similarity measure is adjusted.
Graph Clustering with Local Density-Cut
The basic idea is to envision the graph clustering as a local density-cut problem, where a density-connected tree is first constructed in a local fashion to identify meaningful communities in a graph.
Graph Cuts and Efficient N-D Image Segmentation
This application epitomizes the best features of combinatorial graph cuts methods in vision: global optima, practical efficiency, numerical robustness, ability to fuse a wide range of visual cues and constraints, unrestricted topological properties of segments, and applicability to N-D problems.
Graph Clustering with Density-Cut
This paper introduces a new graph clustering algorithm Dcut from a density point of view, which allows partitioning a graph into multiple densely tight-knit clusters directly and provides a connectivity map of vertices in a graph from a local density perspective.
A new graph-theoretic approach to clustering and segmentation
  • Massimiliano Pavan, M. Pelillo
  • Computer Science, Mathematics
    2003 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2003. Proceedings.
  • 2003
A framework for the image segmentation problem based on a new graph-theoretic formulation of clustering, motivated by the analogies between the intuitive concept of a cluster and that of a dominant set of vertices, which establishes a correspondence between dominant sets and the extrema of a quadratic form over the standard simplex.
A survey of graph theoretical approaches to image segmentation
Data Clustering based on Optimum-Path Forest and Probability Density Function
A graph whose nodes are data samples and whose arcs connect adjacent samples in a feature space is defined, revealing important theoretical relations among relative fuzzy-connected segmentation, morphological reconstructions, watershed transforms, and clustering by influence zones.


Graph theoretic approach to segmentation of MR images
A novel graph theoretic approach to image segmentation is presented, and its application to tissue segmentation in MR images of the human brain is demonstrated. An undirected adjacency graph G is
Graph-Theoretical Methods for Detecting and Describing Gestalt Clusters
  • C. T. Zahn
  • Computer Science
    IEEE Transactions on Computers
  • 1971
A family of graph-theoretical algorithms based on the minimal spanning tree are capable of detecting several kinds of cluster structure in arbitrary point sets; description of the detected clusters
Graph Theoretic Techniques for Cluster Analysis Algorithms
Some applications of graph theory to clustering
Several graphtheoretic criteria are proposed for use within a general clustering paradigm as a means of developing procedures “in between” the extremes of complete-link and single-link hierarchical partitioning.
Graph theoretical clustering based on limited neighbourhood sets
A Graph-Theoretic Approach to Nonparametric Cluster Analysis
This paper presents a noniterative, graph-theoretic approach to nonparametric cluster analysis that is governed by a single-scalar parameter, requires no starting classification, and is capable of determining the number of clusters.
Image segmentation by clustering
The technique does not require training prototypes but operates in an "unsupervised" mode and is based on a mathematical-pattern recognition model, which achieves a maximum value that is postulated to represent an intrinsic number of clusters in the data.
A Dynamic Programming Algorithm for Cluster Analysis
A dynamic programming approach is presented that reduces the amount of redundant transitional calculations implicit in a total enumeration approach to partitioning N entities into M disjoint and nonempty subsets clusters.
Tissue Classification In MR Images Using Hierarchical Segmentation
  • Zhenyu Wu, R. Leahy
  • Computer Science
    1990 IEEE Nuclear Science Symposium Conference Record
  • 1990
A new unsupervised hierarchical segmentation algorithm upon which a tissue classification scheme is based is developed and a novel graph theoretic clustering algorithm is developed to reduce the number of resulting connected regions.
Segmentation of a Thematic Mapper Image Using the Fuzzy c-Means Clusterng Algorthm
Results are presented to show that this two-stage process leads to separation of corn and soybean, and of several minor classes that would otherwise be overwhelmed in any practical one-stage clustering.