Weakly-Supervised Salient Object Detection via Scribble Annotations

TL;DR

This paper introduces a weakly-supervised method for salient object detection using scribble annotations, incorporating edge detection and a scribble boosting scheme to improve boundary accuracy and structure alignment, achieving competitive results.

Contribution

The paper proposes a novel weakly-supervised saliency detection approach utilizing scribble annotations, edge detection, and iterative scribble boosting to enhance boundary localization and structure consistency.

Findings

Outperforms existing weakly-supervised and unsupervised methods.

Achieves results comparable to fully-supervised models.

Introduces a new saliency structure measure aligned with human perception.

Abstract

Compared with laborious pixel-wise dense labeling, it is much easier to label data by scribbles, which only costs 1$\sim$2 seconds to label one image. However, using scribble labels to learn salient object detection has not been explored. In this paper, we propose a weakly-supervised salient object detection model to learn saliency from such annotations. In doing so, we first relabel an existing large-scale salient object detection dataset with scribbles, namely S-DUTS dataset. Since object structure and detail information is not identified by scribbles, directly training with scribble labels will lead to saliency maps of poor boundary localization. To mitigate this problem, we propose an auxiliary edge detection task to localize object edges explicitly, and a gated structure-aware loss to place constraints on the scope of structure to be recovered. Moreover, we design a scribble boosting scheme to iteratively consolidate our scribble annotations, which are then employed as supervision to learn high-quality saliency maps. As existing saliency evaluation metrics neglect to measure structure alignment of the predictions, the saliency map ranking metric may not comply with human perception. We present a new metric, termed saliency structure measure, to measure the structure alignment of the predicted saliency maps, which is more consistent with human perception. Extensive experiments on six benchmark datasets demonstrate that our method not only outperforms existing weakly-supervised/unsupervised methods, but also is on par with several fully-supervised state-of-the-art models. Our code and data is publicly available at https://github.com/JingZhang617/Scribble_Saliency.