A plug-in attribute correction module for generalized zero-shot learning

作者：

Highlights：

• The first work that can correct attributes while preserving the original meaning of each attribute dimension.

• The first plug-in module that aims to facilitate existing approaches and makes ZSL models eligible or even GZSL tasks.

• The corrected attributes can better reflect the visual contexts without losing the integrability in attributes.

摘要

•The first work that can correct attributes while preserving the original meaning of each attribute dimension.•The first plug-in module that aims to facilitate existing approaches and makes ZSL models eligible or even GZSL tasks.•The corrected attributes can better reflect the visual contexts without losing the integrability in attributes.

论文关键词：Generalized zero shot learning (GZSL),Attribute correction,Orthogonal constraint,pluggable module

论文评审过程：Received 2 March 2020, Revised 23 August 2020, Accepted 25 November 2020, Available online 5 December 2020, Version of Record 14 December 2020.

论文官网地址：https://doi.org/10.1016/j.patcog.2020.107767