TANet: Triplet Attention Network for All-In-One Adverse Weather Image Restoration

Hsing-Hua Wang, Fu-Jen Tsai, Yen-Yu Lin, Chia-Wen Lin; Proceedings of the Asian Conference on Computer Vision (ACCV), 2024, pp. 835-851

Adverse weather image restoration aims to remove unwanted degraded artifacts, such as haze, rain, and snow, caused by adverse weather conditions. Existing methods achieve remarkable results in addressing single-weather conditions. However, they face challenges when encountering unpredictable weather conditions, which often happens in real-world scenarios. Although different types of weather conditions exhibit different degraded patterns, they share common characteristics that are highly related and complementary, including occlusions caused by degraded patterns, color distortion, and contrast attenuation due to the scattering of atmospheric particles. Therefore, we focus on leveraging common knowledge across multiple weather conditions to restore images in a unified manner. In this paper, we propose a Triplet Attention Network (TANet) to efficiently and effectively address all-in-one adverse weather image restoration. TANet incorporates three types of attention mechanisms: local pixel-wise attention and global strip-wise attention to address occlusions caused by non-uniform degraded patterns, and global distribution attention to address color distortion and contrast attenuation caused by atmospheric phenomena. By leveraging common knowledge shared among different weather conditions, TANet successfully addresses multiple weather conditions in a unified manner. Experimental results show that TANet efficiently and effectively achieves state-of-the-art performance in all-in-one adverse weather image restoration.