Clustering on Grassmann manifolds via kernel embedding with application to action analysis

Sareh Shirazi; Mehrtash T. Harandi; Conrad Sanderson; Azadeh Alavi; Brian C. Lovell

doi:10.1109/ICIP.2012.6466976

Clustering on Grassmann manifolds via kernel embedding with application to action analysis

Sareh Shirazi, Mehrtash T. Harandi, Conrad Sanderson, Azadeh Alavi, Brian C. Lovell

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Research › peer-review

38 Citations (Scopus)

Abstract

With the aim of improving the clustering of data (such as image sequences) lying on Grassmann manifolds, we propose to embed the manifolds into Reproducing Kernel Hilbert Spaces. To this end, we define a measure of cluster distortion and embed the manifolds such that the distortion is minimised. We show that the optimal solution is a generalised eigenvalue problem that can be solved very efficiently. Experiments on several clustering tasks (including human action clustering) show that in comparison to the recent intrinsic Grassmann k-means algorithm, the proposed approach obtains notable improvements in clustering accuracy, while also being several orders of magnitude faster.

Original language	English
Title of host publication	2012 IEEE International Conference on Image Processing, ICIP 2012 - Proceedings
Pages	781-784
Number of pages	4
DOIs	https://doi.org/10.1109/ICIP.2012.6466976
Publication status	Published - 1 Dec 2012
Externally published	Yes
Event	IEEE International Conference on Image Processing 2012 - Coronado Springs - Disney World, Orlando, United States of America Duration: 30 Sept 2012 → 3 Oct 2012 Conference number: 19th https://ieeexplore.ieee.org/xpl/conhome/6451323/proceeding (Proceedings)

Publication series

Name	Proceedings - International Conference on Image Processing, ICIP
ISSN (Print)	1522-4880

Conference

Conference	IEEE International Conference on Image Processing 2012
Abbreviated title	ICIP 2012
Country/Territory	United States of America
City	Orlando
Period	30/09/12 → 3/10/12
Internet address	https://ieeexplore.ieee.org/xpl/conhome/6451323/proceeding (Proceedings)

Keywords

action analysis
clustering
Grassmann manifolds
kernels
Reproducing Kernel Hilbert Spaces

Access to Document

10.1109/ICIP.2012.6466976

Cite this

Shirazi, S., Harandi, M. T., Sanderson, C., Alavi, A., & Lovell, B. C. (2012). Clustering on Grassmann manifolds via kernel embedding with application to action analysis. In 2012 IEEE International Conference on Image Processing, ICIP 2012 - Proceedings (pp. 781-784). (Proceedings - International Conference on Image Processing, ICIP). https://doi.org/10.1109/ICIP.2012.6466976

@inproceedings{a8a4063978954be88ce82aad1a01be43,

title = "Clustering on Grassmann manifolds via kernel embedding with application to action analysis",

abstract = "With the aim of improving the clustering of data (such as image sequences) lying on Grassmann manifolds, we propose to embed the manifolds into Reproducing Kernel Hilbert Spaces. To this end, we define a measure of cluster distortion and embed the manifolds such that the distortion is minimised. We show that the optimal solution is a generalised eigenvalue problem that can be solved very efficiently. Experiments on several clustering tasks (including human action clustering) show that in comparison to the recent intrinsic Grassmann k-means algorithm, the proposed approach obtains notable improvements in clustering accuracy, while also being several orders of magnitude faster.",

keywords = "action analysis, clustering, Grassmann manifolds, kernels, Reproducing Kernel Hilbert Spaces",

author = "Sareh Shirazi and Harandi, {Mehrtash T.} and Conrad Sanderson and Azadeh Alavi and Lovell, {Brian C.}",

year = "2012",

month = dec,

day = "1",

doi = "10.1109/ICIP.2012.6466976",

language = "English",

isbn = "9781467325332",

series = "Proceedings - International Conference on Image Processing, ICIP",

pages = "781--784",

booktitle = "2012 IEEE International Conference on Image Processing, ICIP 2012 - Proceedings",

note = "IEEE International Conference on Image Processing 2012, ICIP 2012 ; Conference date: 30-09-2012 Through 03-10-2012",

url = "https://ieeexplore.ieee.org/xpl/conhome/6451323/proceeding",

}

Shirazi, S, Harandi, MT, Sanderson, C, Alavi, A & Lovell, BC 2012, Clustering on Grassmann manifolds via kernel embedding with application to action analysis. in 2012 IEEE International Conference on Image Processing, ICIP 2012 - Proceedings. Proceedings - International Conference on Image Processing, ICIP, pp. 781-784, IEEE International Conference on Image Processing 2012, Orlando, Florida, United States of America, 30/09/12. https://doi.org/10.1109/ICIP.2012.6466976

Clustering on Grassmann manifolds via kernel embedding with application to action analysis. / Shirazi, Sareh; Harandi, Mehrtash T.; Sanderson, Conrad et al.
2012 IEEE International Conference on Image Processing, ICIP 2012 - Proceedings. 2012. p. 781-784 (Proceedings - International Conference on Image Processing, ICIP).

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Research › peer-review

TY - GEN

T1 - Clustering on Grassmann manifolds via kernel embedding with application to action analysis

AU - Shirazi, Sareh

AU - Harandi, Mehrtash T.

AU - Sanderson, Conrad

AU - Alavi, Azadeh

AU - Lovell, Brian C.

N1 - Conference code: 19th

PY - 2012/12/1

Y1 - 2012/12/1

N2 - With the aim of improving the clustering of data (such as image sequences) lying on Grassmann manifolds, we propose to embed the manifolds into Reproducing Kernel Hilbert Spaces. To this end, we define a measure of cluster distortion and embed the manifolds such that the distortion is minimised. We show that the optimal solution is a generalised eigenvalue problem that can be solved very efficiently. Experiments on several clustering tasks (including human action clustering) show that in comparison to the recent intrinsic Grassmann k-means algorithm, the proposed approach obtains notable improvements in clustering accuracy, while also being several orders of magnitude faster.

AB - With the aim of improving the clustering of data (such as image sequences) lying on Grassmann manifolds, we propose to embed the manifolds into Reproducing Kernel Hilbert Spaces. To this end, we define a measure of cluster distortion and embed the manifolds such that the distortion is minimised. We show that the optimal solution is a generalised eigenvalue problem that can be solved very efficiently. Experiments on several clustering tasks (including human action clustering) show that in comparison to the recent intrinsic Grassmann k-means algorithm, the proposed approach obtains notable improvements in clustering accuracy, while also being several orders of magnitude faster.

KW - action analysis

KW - clustering

KW - Grassmann manifolds

KW - kernels

KW - Reproducing Kernel Hilbert Spaces

UR - http://www.scopus.com/inward/record.url?scp=84875848721&partnerID=8YFLogxK

U2 - 10.1109/ICIP.2012.6466976

DO - 10.1109/ICIP.2012.6466976

M3 - Conference Paper

AN - SCOPUS:84875848721

SN - 9781467325332

T3 - Proceedings - International Conference on Image Processing, ICIP

SP - 781

EP - 784

BT - 2012 IEEE International Conference on Image Processing, ICIP 2012 - Proceedings

T2 - IEEE International Conference on Image Processing 2012

Y2 - 30 September 2012 through 3 October 2012

ER -

Clustering on Grassmann manifolds via kernel embedding with application to action analysis

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Cite this