Weakly-Supervised Saliency Detection via Salient Object Subitizing

TL;DR

This paper introduces a weakly-supervised saliency detection method using saliency subitizing as supervision, enabling detection of multiple salient objects without pixel-level annotations, and achieves competitive results.

Contribution

The paper proposes a novel weak supervision approach for saliency detection using saliency subitizing, with a two-module model for initial mask generation and iterative refinement.

Findings

Outperforms other weakly-supervised methods

Achieves results comparable to some fully-supervised methods

Validated on five benchmark datasets

Abstract

Salient object detection aims at detecting the most visually distinct objects and producing the corresponding masks. As the cost of pixel-level annotations is high, image tags are usually used as weak supervisions. However, an image tag can only be used to annotate one class of objects. In this paper, we introduce saliency subitizing as the weak supervision since it is class-agnostic. This allows the supervision to be aligned with the property of saliency detection, where the salient objects of an image could be from more than one class. To this end, we propose a model with two modules, Saliency Subitizing Module (SSM) and Saliency Updating Module (SUM). While SSM learns to generate the initial saliency masks using the subitizing information, without the need for any unsupervised methods or some random seeds, SUM helps iteratively refine the generated saliency masks. We conduct extensive experiments on five benchmark datasets. The experimental results show that our method outperforms other weakly-supervised methods and even performs comparably to some fully-supervised methods.