RCSVD: Recursive Clustering with Singular Value Decomposition for Dimension Reduction in Content-Based Retrieval of Large Image/Video Databases

        Efficient indexing in feature space is crucial for many digital library applications. However, the efficiency of spatial indexing techniques usually deteriorates with the increase of dimensionality. A new algorithm, Recursive Clustering with SIngular Value Decomposition (RCSVD) is proposed in this paper for reducing the dimensionality of the feature space. In the proposed algorithm, singular value decomposition and clustering techniques are applied recursively to the feature vectors until the dimensions cannot be further reduced. Performance of the proposed algorithm is evaluated based on a selected set of texture features extracted from satellite images. We report experimental results on the tradeoff between increased storage efficiency (due to reduced dimensionality) and reduced search efficiency to attain the same accuracy in the context of the ubiquitous nearest neighbor search operation. The results show that significant dimension reduction can be achieved by using the proposed algorithm without much impact on efficiency.

By: Alex Thomasian, Vittorio Castelli and Chung-Sheng Li

Published in: RC20704 in


This Research Report is available. This report has been submitted for publication outside of IBM and will probably be copyrighted if accepted for publication. It has been issued as a Research Report for early dissemination of its contents. In view of the transfer of copyright to the outside publisher, its distribution outside of IBM prior to publication should be limited to peer communications and specific requests. After outside publication, requests should be filled only by reprints or legally obtained copies of the article (e.g., payment of royalties). I have read and understand this notice and am a member of the scientific community outside or inside of IBM seeking a single copy only.


Questions about this service can be mailed to reports@us.ibm.com .