Open Access

ARTICLE

Spectral-Semantic Decoupled Rectification Network for Non-Uniform Underwater Image Restoration

Jinshuo Ma, Yang Li^*, Can Guo, Wen Gao, Ruiming Zhang

School of Mathematical Sciences, Dalian Minzu University, Dalian, China

* Corresponding Author: Yang Li. Email:

Computers, Materials & Continua 2026, 89(1), 53 https://doi.org/10.32604/cmc.2026.085999

Received 22 May 2026; Accepted 02 July 2026; Issue published 13 August 2026

Abstract

Underwater image restoration is severely hindered by a tightly coupled degradation process: wavelength-dependent spectral distortion combined with non-uniform, multi-scale spatial scattering. Standard Convolutional Neural Networks (CNNs) and rigid physical priors frequently fail in these dynamic environments, limited by restricted receptive fields, overlooked inter-channel spectral correlations, and severe over-enhancement in photon-starved regions. To break this bottleneck, we propose the Phased Feature Rectification Network (PFR-Net), a decoupled architecture that transforms the ill-posed restoration task into a sequential global spectral calibration and deep semantic refinement paradigm. In the first phase, an efficient Multi-Layer Perceptron (MLP)-based Color Mapping (MLP-CM) module acts as a front-end calibrator to rectify global color casts with minimal parameter overhead. Following this, semantic structures are systematically reconstructed through Dilated Multi-Scale Residual Blocks (DMSRB), which are designed to prevent the truncation of delicate structural gradients and dynamically stabilize content-adaptive feature propagation. To bridge the gap between spatial processing and spectral consistency, we introduce Cross-Covariance Global Attention (XCA), capturing long-range inter-channel dependencies with a linear spatial complexity of

---

Keywords

Spectral-semantic decoupling; adaptive dark channel prior; cross-covariance global attention; non-uniform degradation; underwater image restoration

---

Cite This Article

APA Style

Ma, J., Li, Y., Guo, C., Gao, W., Zhang, R. (2026). Spectral-Semantic Decoupled Rectification Network for Non-Uniform Underwater Image Restoration. Computers, Materials & Continua, 89(1), 53. https://doi.org/10.32604/cmc.2026.085999

Vancouver Style

Ma J, Li Y, Guo C, Gao W, Zhang R. Spectral-Semantic Decoupled Rectification Network for Non-Uniform Underwater Image Restoration. Comput Mater Contin. 2026;89(1):53. https://doi.org/10.32604/cmc.2026.085999

IEEE Style

J. Ma, Y. Li, C. Guo, W. Gao, and R. Zhang, “Spectral-Semantic Decoupled Rectification Network for Non-Uniform Underwater Image Restoration,” Comput. Mater. Contin., vol. 89, no. 1, pp. 53, 2026. https://doi.org/10.32604/cmc.2026.085999

BibTex EndNote RIS