Divide-and-conquer algorithm for creating neighborhood graph for clustering

Virmajoki, Olli; Fränti, Pasi

doi:10.1109/icpr.2004.1334103

Cited by 7 publications

(9 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This reduces the time complexity from to for a single node. The graph construction becomes then the bottleneck but fast approximate variants using k-d tree, divide-and-conquer or projection-based search were considered in [35]. The time complexity of the algorithm can be improved accordingly from to at the cost of slight increase in sse [22], [35].…”

Section: Pairwise Nearest Neighbor Methodsmentioning

confidence: 99%

“…The graph construction becomes then the bottleneck but fast approximate variants using k-d tree, divide-and-conquer or projection-based search were considered in [35]. The time complexity of the algorithm can be improved accordingly from to at the cost of slight increase in sse [22], [35]. All the above variants of agglomerative clustering aim at faster speed except the iterative shrinking which aims at better quality.…”

Section: Pairwise Nearest Neighbor Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Image Segmentation by Pairwise Nearest Neighbor Using Mumford-Shah Model

Shah

Patel

Fränti

2021

Proceedings of CECNet 2021

View full text Add to dashboard Cite

Mumford-Shah model has been used for image segmentation by considering both homogeneity and the shape of the segments jointly. It has been previously optimized by complex mathematical optimization methods like Douglas-Rachford, and a faster but sub-optimal k-means. However, they both suffer from fragmentation caused by non-convex segments. In this paper, we present hierarchical algorithm called Pairwise nearest neighbor (PNN) to optimize the Mumford-Shah model. The merge-based strategy utilizing the connectivity of the pixels prevents isolated fragments to be formed, and in this way, reaches better quality in case of images containing complex shapes.

show abstract

Section: Pairwise Nearest Neighbor Methodsmentioning

confidence: 99%

Section: Pairwise Nearest Neighbor Methodsmentioning

confidence: 99%

Image Segmentation by Pairwise Nearest Neighbor Using Mumford-Shah Model

Shah

Patel

Fränti

2021

Proceedings of CECNet 2021

View full text Add to dashboard Cite

show abstract

“…The purpose of index construction is to organize the dataset 𝑆 with a graph structure. Existing algorithms are generally divided into three index construction strategies: Divide-and-conquer [90], Refinement [29], and Increment [39] (see Appendix E). As Figure 4 (top) show, an algorithm's index construction can be divided into five detailed components (C1-C5).…”

Section: Components For Index Constructionmentioning

confidence: 99%

A Comprehensive Survey and Experimental Comparison of Graph-Based Approximate Nearest Neighbor Search

Wang¹,

Xu²,

Ye³

et al. 2021

Preprint

View full text Add to dashboard Cite

Approximate nearest neighbor search (ANNS) constitutes an important operation in a multitude of applications, including recommendation systems, information retrieval, and pattern recognition. In the past decade, graph-based ANNS algorithms have been the leading paradigm in this domain, with dozens of graph-based ANNS algorithms proposed. Such algorithms aim to provide effective, efficient solutions for retrieving the nearest neighbors for a given query. Nevertheless, these efforts focus on developing and optimizing algorithms with different approaches, so there is a real need for a comprehensive survey about the approaches' relative performance, strengths, and pitfalls. Thus here we provide a thorough comparative analysis and experimental evaluation of 13 representative graph-based ANNS algorithms via a new taxonomy and fine-grained pipeline. We compared each algorithm in a uniform test environment on eight real-world datasets and 12 synthetic datasets with varying sizes and characteristics. Our study yields novel discoveries, offerings several useful principles to improve algorithms. This effort also helped us pinpoint algorithms' working portions, along with rule-of-thumb recommendations about promising research directions and suitable algorithms for practitioners in different fields.

show abstract

“…Several works in connection with the notion of neighbor graph were found in the literature. In [56], the neighbor graph is utilized as a search structure to provide a faster searching method in high-dimensional data sets. In [57], the authors present a number of proximity searching algorithms using the k-nearest neighbor graph as the data structure for searching in metric spaces.…”

Section: Neighbor Graphmentioning

confidence: 99%

“…Several possibilities were proposed in the literature [56][57] [58] for constructing neighbor graphs. One of the common techniques to various neighbor graph construction approaches is based on distances calculation between items.…”

Section: Constructing Neighbor Graphmentioning

confidence: 99%

A pro-active mobility management scheme for publish/subscribe middleware systems

Gaddah¹

View full text Add to dashboard Cite

Divide-and-conquer algorithm for creating neighborhood graph for clustering

Cited by 7 publications

References 9 publications

Image Segmentation by Pairwise Nearest Neighbor Using Mumford-Shah Model

Image Segmentation by Pairwise Nearest Neighbor Using Mumford-Shah Model

A Comprehensive Survey and Experimental Comparison of Graph-Based Approximate Nearest Neighbor Search

A pro-active mobility management scheme for publish/subscribe middleware systems

Contact Info

Product

Resources

About