Informatics Report Series


Report   

EDI-INF-RR-1193


Related Pages

Report (by Number) Index
Report (by Date) Index
Author Index
Institute Index

Home
Title:Multiple Bernoulli Relevance Models for Image and Video Annotation
Authors: S Feng ; R Manmatha ; Victor Lavrenko
Date:Jun 2004
Publication Title:Proceedings of the International Conference on Video and Pattern Reconigtion (CVPR)
Publication Type:Conference Paper Publication Status:Published
Page Nos:II-1002-II
DOI:10.1109/CVPR.2004.1315274
Abstract:
Retrieving images in response to textual queries requires some knowledge of the semantics of the picture. Here we show how we can do both automatic image annotation and retrieval (using one word queries) from images and videos using a multiple Bernoulli relevance model. The model assumes that a training set of images and the specific correspondence between a keyword and an image is not provided. Each image is partitioned into a set of rectangular regions and a real-valued feature vector is computed over these regions. The relevance model is a joint probability distribution of the word annotations and the image feature vectors and is computed using the training set. The word probabilities are estimated using a multiple Bernoulli model and the image feature probabilities using a non-parametric kernel density estimate. The model is then used to annotate images in a test set. We show experiments on both images from a standard Corel data set and a set of video key frames from NIST's video tree. Comparative experiments show that the model performs better than a model based on estimating word probabilities using the popular multinomial disribution. The results also show that our model significantly outperforms previously reported results on the task of image and video annotation.
Links To Paper
No links available
Bibtex format
@InProceedings{EDI-INF-RR-1193,
author = { S Feng and R Manmatha and Victor Lavrenko },
title = {Multiple Bernoulli Relevance Models for Image and Video Annotation},
book title = {Proceedings of the International Conference on Video and Pattern Reconigtion (CVPR)},
year = 2004,
month = {Jun},
pages = {II-1002-II},
doi = {10.1109/CVPR.2004.1315274},
}


Home : Publications : Report 

Please mail <reports@inf.ed.ac.uk> with any changes or corrections.
Unless explicitly stated otherwise, all material is copyright The University of Edinburgh