Learning sparse latent representation and distance metric for image retrieval

The performance of image retrieval depends critically on the semantic representation and the distance function used to estimate the similarity of two images. A good representation should integrate multiple visual and textual (e.g., tag) features and offer a step closer to the true semantics of inter...

Full description

Bibliographic Details
Main Authors: Nguyen, T., Tran, The Truyen, Phung, D., Venkatesh, S.
Format: Conference Paper
Published: 2013
Online Access:http://hdl.handle.net/20.500.11937/19160
_version_ 1848749953575288832
author Nguyen, T.
Tran, The Truyen
Phung, D.
Venkatesh, S.
author_facet Nguyen, T.
Tran, The Truyen
Phung, D.
Venkatesh, S.
author_sort Nguyen, T.
building Curtin Institutional Repository
collection Online Access
description The performance of image retrieval depends critically on the semantic representation and the distance function used to estimate the similarity of two images. A good representation should integrate multiple visual and textual (e.g., tag) features and offer a step closer to the true semantics of interest (e.g., concepts). As the distance function operates on the representation, they are interdependent, and thus should be addressed at the same time. We propose a probabilistic solution to learn both the representation from multiple feature types and modalities and the distance metric from data. The learning is regularised so that the learned representation and information-theoretic metric will (i) preserve the regularities of the visual/textual spaces, (ii) enhance structured sparsity, (iii) encourage small intra-concept distances, and (iv) keep inter-concept images separated. We demonstrate the capacity of our method on the NUS-WIDE data. For the well-studied 13 animal subset, our method outperforms state-of-the-art rivals. On the subset of single-concept images, we gain 79:5% improvement over the standard nearest neighbours approach on the MAP score, and 45.7% on the NDCG.
first_indexed 2025-11-14T07:29:08Z
format Conference Paper
id curtin-20.500.11937-19160
institution Curtin University Malaysia
institution_category Local University
last_indexed 2025-11-14T07:29:08Z
publishDate 2013
recordtype eprints
repository_type Digital Repository
spelling curtin-20.500.11937-191602017-09-13T15:42:22Z Learning sparse latent representation and distance metric for image retrieval Nguyen, T. Tran, The Truyen Phung, D. Venkatesh, S. The performance of image retrieval depends critically on the semantic representation and the distance function used to estimate the similarity of two images. A good representation should integrate multiple visual and textual (e.g., tag) features and offer a step closer to the true semantics of interest (e.g., concepts). As the distance function operates on the representation, they are interdependent, and thus should be addressed at the same time. We propose a probabilistic solution to learn both the representation from multiple feature types and modalities and the distance metric from data. The learning is regularised so that the learned representation and information-theoretic metric will (i) preserve the regularities of the visual/textual spaces, (ii) enhance structured sparsity, (iii) encourage small intra-concept distances, and (iv) keep inter-concept images separated. We demonstrate the capacity of our method on the NUS-WIDE data. For the well-studied 13 animal subset, our method outperforms state-of-the-art rivals. On the subset of single-concept images, we gain 79:5% improvement over the standard nearest neighbours approach on the MAP score, and 45.7% on the NDCG. 2013 Conference Paper http://hdl.handle.net/20.500.11937/19160 10.1109/ICME.2013.6607435 restricted
spellingShingle Nguyen, T.
Tran, The Truyen
Phung, D.
Venkatesh, S.
Learning sparse latent representation and distance metric for image retrieval
title Learning sparse latent representation and distance metric for image retrieval
title_full Learning sparse latent representation and distance metric for image retrieval
title_fullStr Learning sparse latent representation and distance metric for image retrieval
title_full_unstemmed Learning sparse latent representation and distance metric for image retrieval
title_short Learning sparse latent representation and distance metric for image retrieval
title_sort learning sparse latent representation and distance metric for image retrieval
url http://hdl.handle.net/20.500.11937/19160