CARNet: Collaborative Adversarial Resilience for Robust Underwater Image Enhancement and Perception

TL;DR

...

Abstract

Due to the uneven absorption of different light wavelengths in aquatic environments, underwater images suffer from low visibility and clear color deviations. With the advancement of autonomous underwater vehicles, extensive research has been conducted on learning-based underwater enhancement algorithms. These works can generate visually pleasing enhanced images and mitigate the adverse effects of degraded images on subsequent perception tasks. However, learning-based methods are susceptible to the inherent fragility of adversarial attacks, causing significant disruption in enhanced results. In this work, we introduce a collaborative adversarial resilience network, dubbed CARNet, for underwater image enhancement and subsequent detection tasks. Concretely, we first introduce an invertible network with strong perturbation-perceptual abilities to isolate attacks from underwater images, preventing interference with visual quality enhancement and perceptual tasks. Furthermore, an attack pattern discriminator is introduced to adaptively identify and eliminate various types of attacks. Additionally, we propose a bilevel attack optimization strategy to heighten the robustness of the network against different types of attacks under the collaborative adversarial training of vision-driven and perception-driven attacks. Extensive experiments demonstrate that the proposed method outputs visually appealing enhancement images and performs an average 6.71% higher detection mAP than state-of-the-art methods.

Figures (13)

• Figure 1: Overview of the underwater adversarial attack removal process. The proposed method can adaptively adjust the parameter weights of the network by discriminating the different attack patterns. Under the guidance of bilevel attack optimization strategy, the proposed method can effectively improve the effect of underwater image perception under attacks.
• Figure 2: Workflow of the proposed CARNet. In the forward process, the attacked underwater image is transformed into low-resolution image and latent high-frequency component. In the backward process, the high-frequency component is replaced with sampled distribution which is only embedded with clean details to reconstruct enhanced images without attacks.
• Figure 3: The reweighting process from Attack Pattern Discriminator to Dynamic Convolution Layer.
• Figure 4: Enhancement results on UIEBD dataset. The line graph represents the image difference between the clean image and the attack image after enhancement. Smaller differences indicate better robustness.
• Figure 5: Enhancement results on RUIE dataset.