Presentation | 2012-09-03 Nonparametric Bayesian Estimation for Automatic Image Annotation Using Gaussian Mixture Model Yukihiro TSUBOSHITA, Noriji KATO, Masato OKADA, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | Automatic image annotation (AIA) is a process to automatically assign metadata to a digital image in the form of captions or keywords. Here we reported the effort of the improvement with respect to a learning based AIA algorithm using Gaussian mixture model (GMM) as a probabilistic model. In Supervised Multiclass Labeling (SML), which is a conventional method to use GMM, the number of mixed components is identical to all labels. The performance of GMM is known to fully depend on the number of mixed component. Therefore, in the present study, the number of components not being determined in advance, we tried to make the GMM to learn the optimal number of components from given training data. More precisely, we introduced the GMM to Dirichlet process, which is commonly used in the nonparametric Bayesian estimation, as a generating process of mixed components. As the result of evaluation tests using Corel 5K database, which is a standard test collection for image annotation, we found the proposed method exhibited more stable performance than the standard SML. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Automatic image annotation / Machine learning / Gaussian mixture model / Nonparametric Bayesian model |
Paper # | PRMU2012-40,IBISML2012-23 |
Date of Issue |
Conference Information | |
Committee | PRMU |
---|---|
Conference Date | 2012/8/26(1days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | |
Chair | |
Vice Chair | |
Secretary | |
Assistant |
Paper Information | |
Registration To | Pattern Recognition and Media Understanding (PRMU) |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Nonparametric Bayesian Estimation for Automatic Image Annotation Using Gaussian Mixture Model |
Sub Title (in English) | |
Keyword(1) | Automatic image annotation |
Keyword(2) | Machine learning |
Keyword(3) | Gaussian mixture model |
Keyword(4) | Nonparametric Bayesian model |
1st Author's Name | Yukihiro TSUBOSHITA |
1st Author's Affiliation | Corporate Research Group Fuji Xerox Co. Ltd.() |
2nd Author's Name | Noriji KATO |
2nd Author's Affiliation | Corporate Research Group Fuji Xerox Co. Ltd. |
3rd Author's Name | Masato OKADA |
3rd Author's Affiliation | Graduate School of Frontier Sciences The University of Tokyo |
Date | 2012-09-03 |
Paper # | PRMU2012-40,IBISML2012-23 |
Volume (vol) | vol.112 |
Number (no) | 197 |
Page | pp.pp.- |
#Pages | 6 |
Date of Issue |