BCSCN:Reducing Domain Gap through Bézier Curve basis-based Sparse Coding Network for Single-Image Super-Resolution

Wenhao Guo, Peng Lu, Xujun Peng, Zhaoran Zhao, Ji Qiu, XiangTao Dong

Abstract: Single Image Super-Resolution (SISR) is a pivotal challenge in computer vision, aiming to restore high-resolution (HR) images from their low-resolution (LR) counterparts. The presence of diverse degradation kernels creates a significant domain gap, limiting the effective generalization of models in real-world scenarios. This study introduces the Bézier Curve basis-based Sparse Coding Network (BCSCN), a preprocessing network designed to mitigate input distribution discrepancies between the training and testing phases of super-resolution networks. BCSCN achieves this by removing visual defects associated with the degradation kernel in LR images, such as artifacts, residual structures, and noise. Additionally, we propose a set of rewards to guide the search for basis coefficients in BCSCN, enhancing the preservation of main content while eliminating information related to degradation. The experimental results highlight the importance of BCSCN, showcasing its capacity to effectively reduce domain gaps and enhance the generalization of super-resolution networks.

Primary Subject Area: [Content] Media Interpretation

Secondary Subject Area: [Generation] Generative Multimedia

Relevance To Conference: This study focuses on single-image super-resolution (SISR), a longstanding mainstream task in the multimedia domain. Specifically, it addresses a key challenge in SISR—the domain gap caused by varying degradation kernels, which has historically impeded the application of most SISR methods in real-world scenarios. We introduce a novel Bézier Curve-based Sparse Coding Network (BCSCN) that innovatively uses Bézier curve bases for preprocessing degraded images. This approach effectively narrows the distribution gap between different degraded images, enhancing the generalization capability of the foundational super-resolution model across various real-world settings. This method has practical implications for multimedia applications, such as digital video magnification, image restoration, and enhancement of low-resolution multimedia content in realistic environments. It significantly contributes to improving the quality of visual media and enhancing the capability of multimedia systems to effectively process super-resolution tasks in practical scenarios.

Submission Number: 1669