Spatial interest pixels (SIPs): Useful low-level features of visual media data

Qi Li; Jieping Ye; Chandra Kambhamettu

Spatial interest pixels (SIPs): Useful low-level features of visual media data

Qi Li, Jieping Ye, Chandra Kambhamettu

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Visual media data such as an image is the raw data representation for many important applications. The biggest challenge in using visual media data comes from the extremely high dimensionality. We present a comparative study on spatial interest pixels (SIPs), including eight-way (a novel SIP miner), Harris, and Lucas-Kanade, whose extraction is considered as an important step in reducing the dimensionality of visual media data. With extensive case studies, we have shown the usefulness of SIPs as the low-level features of visual media data. A class-preserving dimension reduction algorithm (using GSVD) is applied to further reduce the dimension of feature vectors based on SIPs. The experiments showed its superiority over PCA.

Original language	English (US)
Title of host publication	Proceedings - 3rd IEEE International Conference on Data Mining, ICDM 2003
Pages	163-170
Number of pages	8
State	Published - 2003
Externally published	Yes
Event	3rd IEEE International Conference on Data Mining, ICDM '03 - Melbourne, FL, United States Duration: Nov 19 2003 → Nov 22 2003

Publication series

Name	Proceedings - IEEE International Conference on Data Mining, ICDM
ISSN (Print)	1550-4786

Other

Other	3rd IEEE International Conference on Data Mining, ICDM '03
Country/Territory	United States
City	Melbourne, FL
Period	11/19/03 → 11/22/03

ASJC Scopus subject areas

General Engineering

Cite this

Spatial interest pixels (SIPs): Useful low-level features of visual media data. / Li, Qi; Ye, Jieping; Kambhamettu, Chandra.
Proceedings - 3rd IEEE International Conference on Data Mining, ICDM 2003. 2003. p. 163-170 (Proceedings - IEEE International Conference on Data Mining, ICDM).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Li, Q, Ye, J & Kambhamettu, C 2003, Spatial interest pixels (SIPs): Useful low-level features of visual media data. in Proceedings - 3rd IEEE International Conference on Data Mining, ICDM 2003. Proceedings - IEEE International Conference on Data Mining, ICDM, pp. 163-170, 3rd IEEE International Conference on Data Mining, ICDM '03, Melbourne, FL, United States, 11/19/03.

@inproceedings{3ed74b5fc9484d36b62fee0a75b7026c,

title = "Spatial interest pixels (SIPs): Useful low-level features of visual media data",

abstract = "Visual media data such as an image is the raw data representation for many important applications. The biggest challenge in using visual media data comes from the extremely high dimensionality. We present a comparative study on spatial interest pixels (SIPs), including eight-way (a novel SIP miner), Harris, and Lucas-Kanade, whose extraction is considered as an important step in reducing the dimensionality of visual media data. With extensive case studies, we have shown the usefulness of SIPs as the low-level features of visual media data. A class-preserving dimension reduction algorithm (using GSVD) is applied to further reduce the dimension of feature vectors based on SIPs. The experiments showed its superiority over PCA.",

author = "Qi Li and Jieping Ye and Chandra Kambhamettu",

year = "2003",

language = "English (US)",

isbn = "0769519784",

series = "Proceedings - IEEE International Conference on Data Mining, ICDM",

pages = "163--170",

booktitle = "Proceedings - 3rd IEEE International Conference on Data Mining, ICDM 2003",

note = "3rd IEEE International Conference on Data Mining, ICDM '03 ; Conference date: 19-11-2003 Through 22-11-2003",

}

TY - GEN

T1 - Spatial interest pixels (SIPs)

T2 - 3rd IEEE International Conference on Data Mining, ICDM '03

AU - Li, Qi

AU - Ye, Jieping

AU - Kambhamettu, Chandra

PY - 2003

Y1 - 2003

N2 - Visual media data such as an image is the raw data representation for many important applications. The biggest challenge in using visual media data comes from the extremely high dimensionality. We present a comparative study on spatial interest pixels (SIPs), including eight-way (a novel SIP miner), Harris, and Lucas-Kanade, whose extraction is considered as an important step in reducing the dimensionality of visual media data. With extensive case studies, we have shown the usefulness of SIPs as the low-level features of visual media data. A class-preserving dimension reduction algorithm (using GSVD) is applied to further reduce the dimension of feature vectors based on SIPs. The experiments showed its superiority over PCA.

AB - Visual media data such as an image is the raw data representation for many important applications. The biggest challenge in using visual media data comes from the extremely high dimensionality. We present a comparative study on spatial interest pixels (SIPs), including eight-way (a novel SIP miner), Harris, and Lucas-Kanade, whose extraction is considered as an important step in reducing the dimensionality of visual media data. With extensive case studies, we have shown the usefulness of SIPs as the low-level features of visual media data. A class-preserving dimension reduction algorithm (using GSVD) is applied to further reduce the dimension of feature vectors based on SIPs. The experiments showed its superiority over PCA.

UR - http://www.scopus.com/inward/record.url?scp=26844550244&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=26844550244&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:26844550244

SN - 0769519784

SN - 9780769519784

T3 - Proceedings - IEEE International Conference on Data Mining, ICDM

SP - 163

EP - 170

BT - Proceedings - 3rd IEEE International Conference on Data Mining, ICDM 2003

Y2 - 19 November 2003 through 22 November 2003

ER -

Spatial interest pixels (SIPs): Useful low-level features of visual media data

Abstract

Publication series

Other

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this