Graph Degree Linkage: Agglomerative Clustering on a Directed Graph

@inproceedings{Zhang2012GraphDL,
  title={Graph Degree Linkage: Agglomerative Clustering on a Directed Graph},
  author={Wayne Zhang and Xiaogang Wang and Deli Zhao and Xiaoou Tang},
  booktitle={ECCV},
  year={2012}
}
This paper proposes a simple but effective graph-based agglomerative algorithm, for clustering high-dimensional data. We explore the different roles of two fundamental concepts in graph theory, indegree and outdegree, in the context of clustering. The average indegree reflects the density near a sample, and the average outdegree characterizes the local geometry around a sample. Based on such insights, we define the affinity measure of clusters via the product of average indegree and average… 
Homophilic Clustering by Locally Asymmetric Geometry
TLDR
A novel algorithm to address the challenge of detecting clusters from heavy noise using the in-degree homophily, and develops a simple efficient algorithm of linear space complexity to cluster noisy data.
Adapting k-means for graph clustering
TLDR
Two new algorithms for clustering graphs and networks derived directly from the k-means algorithm are proposed, one of which outperforms eight other state-of-the-art methods and the other gradually improves on the results of the K-algorithm to find new and potentially better local optima.
Data Clustering Based on Community Structure in Mutual k-Nearest Neighbor Graph
TLDR
A novel data clustering algorithm that uses the authentic scores-a metric that measures the strength of an edge in a social network graph-to rank all the edges in the MKNN graph and collapses the graph into components to find clusters of arbitrary shape is proposed.
Sampling Clustering
TLDR
An efficient linear-time graph-based divisive cluster analysis approach called Sampling Clustering that constructs a lite informative dendrogram by recursively dividing a graph into subgraphs that outperform state-of-the-art clustering algorithms with significantly less computing resource requirements.
Object co-segmentation based on directed graph clustering
TLDR
A new algorithm to segment multiple common objects from a group of images by directed graph clustering and prior propagation that can achieve larger accuracy compared with several state-of-arts co-segmentation methods.
Agglomerative clustering and collectiveness measure via exponent generating function
TLDR
A novel agglomerative clustering method is proposed by utilizing the path integral to define the affinity measure and the proposed path integral descriptor can be regard as the collectiveness measure of a set, which can be a moving system such as human crowd, sheep herd and so on.
Density Peak Clustering Based on Cumulative Nearest Neighbors Degree and Micro Cluster Merging
TLDR
A new clustering method, density peak clustering based on cumulative nearest neighbors degree and micro cluster merging, which improves the DPC algorithm in two ways, the one is that the method defines a new local density to solve the defect of the D PC algorithm and the other one is the graph degree linkage is combined with theDPC to alleviate the problem of distribution errors.
EDDPC: An Efficient Distributed Density Peaks Clustering Algorithm
TLDR
An efficient distributed density peaks clustering algorithm, EDDPC, is proposed, which leverages Voronoi diagram and careful data replication/filtering to reduce huge amount of useless distance measurement cost and data shuffle cost and the results show that the algorithm can improve the performance significantly compared to naive MapReduce implementation.
Constrained Directed Graph Clustering and Segmentation Propagation for Multiple Foregrounds Cosegmentation
TLDR
A new constrained directed graph clustering (DGC) method and segmentation propagation method for the multiple foreground cosegmentation with the perspective of classification and propagation, where the classification is used to obtain the object prior of each class and the propagation isused to propagate the prior to all images.
...
...

References

SHOWING 1-10 OF 31 REFERENCES
Cyclizing Clusters via Zeta Function of a Graph
TLDR
An algorithm named Zeta l-links (Zell) is developed which consists of two parts: Zeta merging with a similarity graph and an initial set of small clusters derived from local l- links of samples, to structurize a cluster using cycles in the associated subgraph.
Fast Agglomerative Clustering Using a k-Nearest Neighbor Graph
TLDR
A fast agglomerative clustering method using an approximate nearest neighbor graph for reducing the number of distance calculations and a relatively small neighborhood size is sufficient to maintain the quality close to that of the full search.
Chameleon: Hierarchical Clustering Using Dynamic Modeling
TLDR
Chameleon's key feature is that it accounts for both interconnectivity and closeness in identifying the most similar pair of clusters, which is important for dealing with highly variable clusters.
Finding Clusters of Different Sizes, Shapes, and Densities in Noisy, High Dimensional Data
TLDR
A novel clustering technique that addresses problems with varying densities and high dimensionality, while the use of core points handles problems with shape and size, and a number of optimizations that allow the algorithm to handle large data sets are discussed.
Density-Based Clustering in Spatial Databases: The Algorithm GDBSCAN and Its Applications
TLDR
The generalized algorithm DBSCAN can cluster point objects as well as spatially extended objects according to both, their spatial and their nonspatial attributes, and four applications using 2D points (astronomy, 3D points,biology, 5D points and 2D polygons) are presented, demonstrating the applicability of GDBSCAN to real-world problems.
Normalized cuts and image segmentation
  • Jianbo Shi, J. Malik
  • Computer Science
    Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition
  • 1997
TLDR
This work treats image segmentation as a graph partitioning problem and proposes a novel global criterion, the normalized cut, for segmenting the graph, which measures both the total dissimilarity between the different groups as well as the total similarity within the groups.
Common visual pattern discovery via spatially coherent correspondences
  • Hairong Liu, Shuicheng Yan
  • Computer Science
    2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition
  • 2010
TLDR
This paper investigates how to discover all common visual patterns within two sets of feature points and finds all large local maxima of a quadratic function on simplex, using replicator equation and through a systematic way of initialization.
Learning from labeled and unlabeled data on a directed graph
TLDR
A general framework for learning from labeled and unlabeled data on a directed graph in which the structure of the graph including the directionality of the edges is considered, which generalizes the spectral clustering approach for undirected graphs.
Feature correspondence and deformable object matching via agglomerative correspondence clustering
TLDR
The algorithm handles significant amount of outliers and deformation as well as multiple clusters, thus enabling simultaneous feature matching and clustering from real-world image pairs with significant clutter and multiple deformable objects.
A spectral technique for correspondence problems using pairwise constraints
  • M. Leordeanu, M. Hebert
  • Computer Science
    Tenth IEEE International Conference on Computer Vision (ICCV'05) Volume 1
  • 2005
TLDR
An efficient spectral method for finding consistent correspondences between two sets of features by using the principal eigenvector of M and imposing the mapping constraints required by the overall correspondence mapping.
...
...