An image similarity descriptor for classification tasks

We develop an image similarity descriptor for an image pair, based on deep features. The development consists of two parts - selecting the deep layer whose features are to be included in the descriptor, and a representation of the similarity between the images in the pair. The selection of the deep...

Full description

Saved in:
Bibliographic Details
Main Authors: Wang, Liangliang, Rajan, Deepu
Other Authors: School of Computer Science and Engineering
Format: Article
Language:English
Published: 2022
Subjects:
Online Access:https://hdl.handle.net/10356/161045
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
id sg-ntu-dr.10356-161045
record_format dspace
spelling sg-ntu-dr.10356-1610452022-08-12T06:39:33Z An image similarity descriptor for classification tasks Wang, Liangliang Rajan, Deepu School of Computer Science and Engineering Media & Interactive Computing Lab Engineering::Computer science and engineering Image Similarity Similarity Representation We develop an image similarity descriptor for an image pair, based on deep features. The development consists of two parts - selecting the deep layer whose features are to be included in the descriptor, and a representation of the similarity between the images in the pair. The selection of the deep layer follows a sparse representation of the feature maps followed by multi-output support vector regression. The similarity representation is based on a novel correlation between the histograms of the feature maps of the two images. Experiments to demonstrate the effectiveness of the proposed descriptor are carried out on four applications that can be cast as classification tasks. 2022-08-12T06:39:33Z 2022-08-12T06:39:33Z 2020 Journal Article Wang, L. & Rajan, D. (2020). An image similarity descriptor for classification tasks. Journal of Visual Communication and Image Representation, 71, 102847-. https://dx.doi.org/10.1016/j.jvcir.2020.102847 1047-3203 https://hdl.handle.net/10356/161045 10.1016/j.jvcir.2020.102847 2-s2.0-85088361021 71 102847 en Journal of Visual Communication and Image Representation © 2020 Elsevier Inc. All rights reserved.
institution Nanyang Technological University
building NTU Library
continent Asia
country Singapore
Singapore
content_provider NTU Library
collection DR-NTU
language English
topic Engineering::Computer science and engineering
Image Similarity
Similarity Representation
spellingShingle Engineering::Computer science and engineering
Image Similarity
Similarity Representation
Wang, Liangliang
Rajan, Deepu
An image similarity descriptor for classification tasks
description We develop an image similarity descriptor for an image pair, based on deep features. The development consists of two parts - selecting the deep layer whose features are to be included in the descriptor, and a representation of the similarity between the images in the pair. The selection of the deep layer follows a sparse representation of the feature maps followed by multi-output support vector regression. The similarity representation is based on a novel correlation between the histograms of the feature maps of the two images. Experiments to demonstrate the effectiveness of the proposed descriptor are carried out on four applications that can be cast as classification tasks.
author2 School of Computer Science and Engineering
author_facet School of Computer Science and Engineering
Wang, Liangliang
Rajan, Deepu
format Article
author Wang, Liangliang
Rajan, Deepu
author_sort Wang, Liangliang
title An image similarity descriptor for classification tasks
title_short An image similarity descriptor for classification tasks
title_full An image similarity descriptor for classification tasks
title_fullStr An image similarity descriptor for classification tasks
title_full_unstemmed An image similarity descriptor for classification tasks
title_sort image similarity descriptor for classification tasks
publishDate 2022
url https://hdl.handle.net/10356/161045
_version_ 1743119483984150528