A scaled-MST-based clustering algorithm and application on image segmentation

Li, Jia; Wang, Xiaochun; Wang, Xiali

doi:10.1007/s10844-019-00572-x

Cited by 8 publications

(10 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…First, the proposed Best Friend Clustering is evaluated with three classic 2D datasets. Results of KMeans [50], AHC [39] and MSTC [34] methods are also presented as different baselines for comparison. Then we turn to the evaluation on convergence, accuracy, and scalability for our regression library with five real large-scale datasets.…”

Section: Data Organizationmentioning

confidence: 99%

“…Since DCKRR [53] and BKRR2 [49] are two closely related papers, they are employed as two baselines in this paper. Moreover, we alternate the clustering methods in BKRR2 with AHC [39] and MSTC [34] as AHCKRR and MSTKRR respectively. Thus the experimental configurations for parallel regression cover the whole spectrum of representative clustering methods: BFCKRR (Best Friend Clustering), BKRR2 (K-Means), AHCKRR (Agglomerative Hierarchical Clustering), and MSTKRR (Minimum Spanning Tree Clustering).…”

Section: Data Organizationmentioning

confidence: 99%

See 1 more Smart Citation

An Accurate and Efficient Large-scale Regression Method through Best Friend Clustering

Li¹,

Liu²,

Zhang³

et al. 2021

Preprint

View full text Add to dashboard Cite

As the data size in Machine Learning fields grows exponentially, it is inevitable to accelerate the computation by utilizing the evergrowing large number of available cores provided by high-performance computing hardware. However, existing parallel methods for clustering or regression often suffer from problems of low accuracy, slow convergence, and complex hyperparameter-tuning. Furthermore, the parallel efficiency is usually difficult to improve while striking a balance between preserving model properties and partitioning computing workloads on distributed systems. In this paper, we propose a novel and simple data structure capturing the most important information among data samples. It has several advantageous properties supporting a hierarchical clustering strategy that is irrelevant to the hardware parallelism, well-defined metrics for determining optimal clustering, balanced partition for maintaining the compactness property, and efficient parallelization for accelerating computation phases. Then we combine the clustering with regression techniques as a parallel library and utilize a hybrid structure of data and model parallelism to make predictions. Experiments illustrate that our library obtains remarkable performance on convergence, accuracy, and scalability.

show abstract

Section: Data Organizationmentioning

confidence: 99%

Section: Data Organizationmentioning

confidence: 99%

An Accurate and Efficient Large-scale Regression Method through Best Friend Clustering

Li¹,

Liu²,

Zhang³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Aiming at the scale defect of the local filter, identifying global characteristics and connecting global information is a potentially effective means. Referece [ 31 ] indicates that the image segmentation process can be handled as a clustering problem and the MST can preserve the connectivity of the image graph and can provide a link to all nodes at a minimum total edge cost during clustering, which is verified in follow-up researches [ 32 , 33 ]. An efficient MST based global filtering method for image matching is first proposed in [ 32 ].…”

Section: Related Workmentioning

confidence: 99%

“…An efficient MST based global filtering method for image matching is first proposed in [ 32 ]. Furthermore, the improvements have been made to address limitations for data sets with different density distribution in [ 33 ]. Besides, compared with the uncertainty of connectivity and the complexity of solving the non-deterministic polynomial (NP)-hard problem in the normalized cutting method [ 34 , 35 ], the MST can preserve all the important edge information without requiring any closing or connection of the edges and its pixels spatial relationship provides the possibility of fusion with local filtering algorithms.…”

Section: Related Workmentioning

confidence: 99%

“…The MST structure [ 33 ] for an image which involves pixel spatial distance and color/intensity difference provides a natural measure of global pixel similarity and has the potential to solve the problem (3). MST for an image regards an image as a 4-connected, undirected graph

, where

is the vertex formed by all the pixels in the image and

is the edge constituted by all edges between connected pixels.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

A Novel Post-Processing Method Based on a Weighted Composite Filter for Enhancing Semantic Segmentation Results

Cheng

Liu

2020

Sensors

View full text Add to dashboard Cite

Image semantic segmentation is one of the key problems in computer vision. Despite the enormous advances in applications, almost all the image semantic segmentation algorithms fail to achieve satisfactory segmentation results due to lack of sensitivity to details, or difficulty in evaluating the global similarity of pixels, or both. Posting-processing enhancement methods, as the outstandingly crucial means to ameliorate the above-mentioned inherent flaws of algorithms, are almost based on conditional random fields (CRFs). Inspired by CRFs, this paper proposes a novel post-processing enhancement framework with theoretical simplicity from the perspective of filtering, and a new weighted composite filter (WCF) is designed to enhance the segmentation masks in a unified framework. First, by adjusting the weight ratio, the WCF is decomposed into a local part and a global part. Secondly, a guided image filter is designed as the local filter, which can restore boundary information to present necessary details. Moreover, a minimum spanning tree (MST)-based filter is designed as the global filter to provide a natural measure of global pixel similarity for image matching. Thirdly, a unified post-processing enhancement framework, including selection and normalization, WCF and argmax, is designed. Finally, the effectiveness and superiority of the proposed method for enhancement, as well as its range of applications, are verified through experiments.

show abstract

A minimum spanning tree based partitioning and merging technique for clustering heterogeneous data sets

Mishra

2020

J Intell Inf Syst

View full text Add to dashboard Cite

A scaled-MST-based clustering algorithm and application on image segmentation

Cited by 8 publications

References 31 publications

An Accurate and Efficient Large-scale Regression Method through Best Friend Clustering

An Accurate and Efficient Large-scale Regression Method through Best Friend Clustering

A Novel Post-Processing Method Based on a Weighted Composite Filter for Enhancing Semantic Segmentation Results

A minimum spanning tree based partitioning and merging technique for clustering heterogeneous data sets

Contact Info

Product

Resources

About