On Implicit Attribute Localization for Generalized Zero-Shot Learning

TL;DR

This paper reveals that standard zero-shot learning models can implicitly localize attributes without explicit mechanisms, and introduces SELAR, a simple method that enhances attribute localization and achieves competitive results.

Contribution

The paper demonstrates implicit attribute localization in common ZSL models and proposes SELAR, a straightforward approach that improves GZSL performance without complex modules.

Findings

Implicit attribute localization is present in common ZSL backbones.

SELAR enhances attribute localization and achieves competitive GZSL results.

The method provides a simple yet effective baseline for future GZSL research.

Abstract

Zero-shot learning (ZSL) aims to discriminate images from unseen classes by exploiting relations to seen classes via their attribute-based descriptions. Since attributes are often related to specific parts of objects, many recent works focus on discovering discriminative regions. However, these methods usually require additional complex part detection modules or attention mechanisms. In this paper, 1) we show that common ZSL backbones (without explicit attention nor part detection) can implicitly localize attributes, yet this property is not exploited. 2) Exploiting it, we then propose SELAR, a simple method that further encourages attribute localization, surprisingly achieving very competitive generalized ZSL (GZSL) performance when compared with more complex state-of-the-art methods. Our findings provide useful insight for designing future GZSL methods, and SELAR provides an easy to implement yet strong baseline.