Enhancing Underwater Images via Asymmetric Multi-Scale Invertible Networks

Yuhui Quan, Xiaoheng Tan, Yan Huang, Yong Xu, Hui Ji

Abstract: Underwater images, often plagued by complex degradation, pose significant challenges for image enhancement. To address these challenges, the paper redefines underwater image enhancement as an image decomposition problem and proposes a deep invertible neural network (INN) that accurately predicts both the latent image and the degradation effects. Instead of using an explicit formation model to describe the degradation process, the INN adheres to the constraints of the image decomposition model, providing necessary regularization for model training, particularly in the absence of supervision on degradation effects. Taking into account the diverse scales of degradation factors, the INN is structured on a multi-scale basis to effectively manage the varied scales of degradation factors. Moreover, the INN incorporates several asymmetric design elements that are specifically optimized for the decomposition model and the unique physics of underwater imaging. Comprehensive experiments show that our approach provides significant performance improvement over existing methods.

Primary Subject Area: [Experience] Multimedia Applications

Secondary Subject Area: [Content] Media Interpretation

Relevance To Conference: Multimedia platforms, such as websites, documentaries, educational videos, and presentations, have gained popularity for communicating the discoveries and adventures of underwater exploration to broader audiences. They effectively showcase the allure, difficulties, and importance of underwater environments, fostering public awareness about the significance of underwater exploration and conservation efforts. Given the inherent challenges of underwater photography, which often faces significant degradation effects, the enhancement of underwater images becomes imperative for achieving high-quality results. Therefore, underwater image enhancement stands as a vital technique in multimedia processing. The invertible decomposition method proposed in this paper not only addresses these challenges but also offers insights that may inspire advancements in other image processing techniques for multimedia applications.

Submission Number: 2548