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

Authors:
Qi Li;Jieping Ye;Chandra Kambhamettu
Affiliations:
Computer Science & Engineering, Arizona State University, Tempe, USA 85281;Computer Science & Engineering, Arizona State University, Tempe, USA 85281;Video/Image Modeling and Synthesis Lab Computer Information & Sciences, University of Delaware, Newark, USA 19716
Venue:
Multimedia Tools and Applications
Year:
2006

Citing 0
Cited 1

A geometric data structure applicable to image mining and retrieval

ICIAR'10 Proceedings of the 7th international conference on Image Analysis and Recognition - Volume Part I

Quantified Score

Hi-index	0.00

Visualization

Abstract

Visual media data such as an image is the raw data representation for many important applications. Reducing the dimensionality of raw visual media data is desirable since high dimensionality degrades not only the effectiveness but also the efficiency of visual recognition algorithms. We present a comparative study on spatial interest pixels (SIPs), including eight-way (a novel SIP detector), Harris, and Lucas驴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 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.