Title

Region Graph Embedding Network for Zero-Shot Learning

Source of Publication

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

Abstract

© 2020, Springer Nature Switzerland AG. Most of the existing Zero-Shot Learning (ZSL) approaches learn direct embeddings from global features or image parts (regions) to the semantic space, which, however, fail to capture the appearance relationships between different local regions within a single image. In this paper, to model the relations among local image regions, we incorporate the region-based relation reasoning into ZSL. Our method, termed as Region Graph Embedding Network (RGEN), is trained end-to-end from raw image data. Specifically, RGEN consists of two branches: the Constrained Part Attention (CPA) branch and the Parts Relation Reasoning (PRR) branch. CPA branch is built upon attention and produces the image regions. To exploit the progressive interactions among these regions, we represent them as a region graph, on which the parts relation reasoning is performed with graph convolutions, thus leading to our PRR branch. To train our model, we introduce both a transfer loss and a balance loss to contrast class similarities and pursue the maximum response consistency among seen and unseen outputs, respectively. Extensive experiments on four datasets well validate the effectiveness of the proposed method under both ZSL and generalized ZSL settings.

Document Type

Conference Proceeding

ISBN

9783030585471

First Page

562

Last Page

580

Publication Date

1-1-2020

DOI

10.1007/978-3-030-58548-8_33

Scopus ID

85097437759

This document is currently not available here.

Share

COinS