WXSOD: A Benchmark for Robust Salient Object Detection in Adverse Weather Conditions

TL;DR

This work tackles robust salient object detection under adverse weather by introducing the WXSOD dataset, a large-scale RGB benchmark with synthetic and real weather noise labels. It also proposes WFANet, a dual-branch network that learns weather-specific noise representations in a weather-prediction branch and fuses them with semantic saliency features in a saliency-detection branch. Experiments against 17 SOD methods demonstrate WFANet achieves superior accuracy on both synthetic and real weather-noise data, highlighting the practical value of weather-aware representation learning. Together, WXSOD and WFANet provide a rigorous benchmark and a strong baseline for RGB SOD in adverse conditions, enabling targeted improvements in robustness and generalization.

Abstract

Salient object detection (SOD) in complex environments remains a challenging research topic. Most existing methods perform well in natural scenes with negligible noise, and tend to leverage multi-modal information (e.g., depth and infrared) to enhance accuracy. However, few studies are concerned with the damage of weather noise on SOD performance due to the lack of dataset with pixel-wise annotations. To bridge this gap, this paper introduces a novel Weather-eXtended Salient Object Detection (WXSOD) dataset. It consists of 14,945 RGB images with diverse weather noise, along with the corresponding ground truth annotations and weather labels. To verify algorithm generalization, WXSOD contains two test sets, i.e., a synthesized test set and a real test set. The former is generated by adding weather noise to clean images, while the latter contains real-world weather noise. Based on WXSOD, we propose an efficient baseline, termed Weather-aware Feature Aggregation Network (WFANet), which adopts a fully supervised two-branch architecture. Specifically, the weather prediction branch mines weather-related deep features, while the saliency detection branch fuses semantic features extracted from the backbone with weather features for SOD. Comprehensive comparisons against 17 SOD methods shows that our WFANet achieves superior performance on WXSOD. The code and benchmark results will be made publicly available at https://github.com/C-water/WXSOD

Figures (14)

• Figure 1: Common weather noise interferes with salient objects. Green curve represents the edge of the salient target predicted by MEANetliang2024meanet, while red curve represents the ground-truth.
• Figure 2: Examples of WXSOD dataset. The first row shows scenes with different sizes of salient objects, and the second row shows scenes with different numbers of salient objects.
• Figure 3: Examples of synthesized test set (a) and real test set (b) in the WXSOD dataset.
• Figure 4: Visualization of various datasets using t-SNE.
• Figure 5: The overall architecture of the proposed WFANet, comprising two branches: a weather prediction branch dedicated to learning noise-related features, and a saliency detection branch for SOD.