Please use this identifier to cite or link to this item: http://hdl.handle.net/10397/11932
Title: Dimensionality reduction for descriptor generation in rushes editing
Authors: Liu, Y
Liu, Y 
Chan, KCC 
Keywords: Multimedia computing
Pattern clustering
Video signal processing
Issue Date: 2008
Publisher: IEEE
Source: 2008 IEEE International Conference on Semantic Computing, 4-7 August 2008, Santa Clara, CA, p. 104-111 How to cite?
Abstract: Rushes editing, which enables the computer to edit the film like a professional film cutter based on the noisy and redundant footage, is an active topic in multimedia semantic analysis. The most critical problem of rushes editing is how to generate an effective, efficient, and robust descriptor for the footage content analysis. This paper proposes a novel non-linear dimensionality reduction algorithm called Multi-Layer Isometric Feature Mapping (ML-Isomap)for automatic descriptor generation. First, a K-nearest Neighbor Based Clustering (KNBC) algorithm is utilized to partition the high-dimensional data points into a set of data blocks. Second, intra-cluster graphs are constructed based on the individual character of each data block to build the basic layer for the ML-Isomap. Third, the inter-cluster graph is constructed by analyzing the interrelation among these isolated data blocks to build the hyper-layers for the ML-Isomap. Finally, all the data points are mapped into the unique low-dimensional feature space by keeping the corresponding relations of the multiple layers in the high-dimensional feature space to the greatest extent. The comparative experiments on synthetic data as well as the real rushes editing tasks demonstrate that the proposed algorithm can generate the effective descriptor with much lower dimensions for the semantic video analysis.
URI: http://hdl.handle.net/10397/11932
ISBN: 978-0-7695-3279-0
978-0-7695-3279-0 (E-ISBN)
DOI: 10.1109/ICSC.2008.90
Appears in Collections:Conference Paper

Access
View full-text via PolyU eLinks SFX Query
Show full item record

Page view(s)

20
Last Week
0
Last month
Checked on Jun 25, 2017

Google ScholarTM

Check

Altmetric



Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.