Toggle Main Menu Toggle Search

Open Access padlockePrints

Attribute relaxation from class level to instance level for zero-shot learning

Lookup NU author(s): Dr Yang Long



This is the authors' accepted manuscript of an article that has been published in its final definitive form by Institution of Engineering and Technology, 2018.

For re-use rights please refer to the publisher's terms and conditions.


© 2018 Institution of Engineering and Technology. All rights reserved. Conventional zero-shot learning (ZSL) methods usually use class-level attribute, which corresponds to a batch of images of same category. This setting is not reasonable since the images even though belong to same category still have variances in their attribute items. To alleviate this phenomenon, the authors propose a novel method namely attribute relaxation (AR) to extend attributes from class level to instance level by adding a small variance matrix, which is more reasonable than traditional ZSL methods such as Semantic AutoEncoder that projects features from multi to one. Extensive experiments on four popular datasets show that AR can significantly improve the method using only class-level attributes, and verifies that AR can make the projected features in attribute space more discriminative.

Publication metadata

Author(s): Zhang H, Long Y, Zhao C

Publication type: Article

Publication status: Published

Journal: Electronics Letters

Year: 2018

Volume: 54

Issue: 20

Pages: 1170-1172

Print publication date: 04/10/2018

Acceptance date: 02/04/2018

Date deposited: 10/01/2019

ISSN (print): 0013-5194

ISSN (electronic): 1350-911X

Publisher: Institution of Engineering and Technology


DOI: 10.1049/el.2018.5027


Altmetrics provided by Altmetric