Learning Optimal Compact Codebook for Efficient Object Categorization
Teng Li
Tao Mei
In So Kweon
Korea Adv. Inst. of Sci. & Technol., Daejeon;
This paper appears in: Applications of Computer Vision, 2008. WACV 2008. IEEE Workshop on
Publication Date: 7-9 Jan. 2008
On page(s): 1-6
Location: Copper Mountain, CO,
ISSN: 1550-5790
ISBN: 978-1-4244-1913-5
INSPEC Accession Number: 10067265
Digital Object Identifier: 10.1109/WACV.2008.4544027
Current Version Published: 2008-06-17
Abstract
Representation of images using the distribution of local features on a visual codebook is an effective method for object categorization. Typically, discriminative capability of the codebook can lead to a better performance. However, conventional methods usually use clustering algorithms to learn codebooks without considering this. This paper presents a novel approach of learning optimal compact codebooks by selecting a subset of discriminative codes from a large codebook. Firstly, the Gaussian models of object categories based on a single code are learned from the distribution of local features within each image. Then two discriminative criteria, i.e. likelihood ratio and Fisher, are introduced to evaluate how each code contributes to the categorization. We evaluate the optimal codebooks constructed by these two criteria on Caltech-4 dataset, and report superior performance of object categorization compared with traditional K-means method with the same size of codebook.
Index
Terms
Available to subscribers and IEEE members.
References
Available to subscribers and IEEE members.
Citing Documents
Available to subscribers and IEEE members.
Cart
