SINET: Sparsity-driven Interpretable Neural Network for Underwater Image Enhancement

2024-09-02

IEEE International Conference on Acoustics, Speech, and Signal Processing

AI Summary

To address severe degradation and insufficient interpretability in underwater image enhancement, this paper proposes a channel-specific convolutional sparse coding (CCSC)-driven interpretable neural network. Methodologically, it is the first to embed ℓ₁-regularized convolutional sparse coding into a deep network via algorithm unrolling, employing three Sparse Feature Estimation Blocks (SFEBs) to separately model physically salient features for the RGB channels—enabling end-to-end optimization while preserving strong interpretability. The key contributions are: (1) unifying model lightweighting (reducing computational complexity by 3873×) with physically traceable feature representations; and (2) achieving state-of-the-art performance on standard underwater image enhancement benchmarks, with a PSNR gain of +1.05 dB over prior methods. This work establishes a new paradigm for marine visual perception that jointly ensures high fidelity and trustworthy interpretability.

Technology Category

• Computer Vision: Interpretability, Explainability, and Transparency
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Machine Learning: Calibration & Uncertainty Quantification

Application Category

• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Improving the quality of underwater images is essential for advancing marine research and technology. This work introduces a sparsity-driven interpretable neural network (SINET) for the underwater image enhancement (UIE) task. Unlike pure deep learning methods, our network architecture is based on a novel channel-specific convolutional sparse coding (CCSC) model, ensuring good interpretability of the underlying image enhancement process. The key feature of SINET is that it estimates the salient features from the three color channels using three sparse feature estimation blocks (SFEBs). The architecture of SFEB is designed by unrolling an iterative algorithm for solving the $ell_1$ regularized convolutional sparse coding (CSC) problem. Our experiments show that SINET surpasses state-of-the-art PSNR value by $1.05$ dB with $3873$ times lower computational complexity. Code can be found at: https://github.com/gargi884/SINET-UIE/tree/main.

Problem

• Enhances underwater image quality for marine research.
• Introduces SINET with sparsity-driven interpretable neural network.
• Reduces computational complexity while improving PSNR values.

Innovation

• Sparsity-driven neural network for underwater images
• Channel-specific convolutional sparse coding model
• Three sparse feature estimation blocks for color channels