DSDNet: Raw Domain Demoir'eing via Dual Color-Space Synergy

2025-04-22

AI Summary

Moiré pattern degradation in smartphone-captured screen images remains challenging: existing sRGB-domain methods suffer from irreversible information loss, while two-stage raw-domain approaches face bottlenecks in information flow and inefficient inference. This paper proposes the first single-stage raw-domain Moiré removal framework, leveraging synergistic modeling of raw and YCbCr representations to achieve high-fidelity Moiré suppression while preserving luminance accuracy and chrominance fidelity. Key innovations include: (1) a Synergic Attention with Dynamic Modulation (SADM) module for cross-domain feature guidance; and (2) a Luminance-Chrominance Adaptive Transformer (LCAT) that decouples and adaptively optimizes luminance and chrominance representations. Extensive experiments demonstrate state-of-the-art performance—outperforming prior methods in PSNR and SSIM—and 2.4× faster inference than the second-best approach, achieving an optimal balance of accuracy and practical efficiency.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Mixture of Experts (MoE)
• Application Domains: Mobility, Driving & Flight

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

With the rapid advancement of mobile imaging, capturing screens using smartphones has become a prevalent practice in distance learning and conference recording. However, moir'e artifacts, caused by frequency aliasing between display screens and camera sensors, are further amplified by the image signal processing pipeline, leading to severe visual degradation. Existing sRGB domain demoir'eing methods struggle with irreversible information loss, while recent two-stage raw domain approaches suffer from information bottlenecks and inference inefficiency. To address these limitations, we propose a single-stage raw domain demoir'eing framework, Dual-Stream Demoir'eing Network (DSDNet), which leverages the synergy of raw and YCbCr images to remove moir'e while preserving luminance and color fidelity. Specifically, to guide luminance correction and moir'e removal, we design a raw-to-YCbCr mapping pipeline and introduce the Synergic Attention with Dynamic Modulation (SADM) module. This module enriches the raw-to-sRGB conversion with cross-domain contextual features. Furthermore, to better guide color fidelity, we develop a Luminance-Chrominance Adaptive Transformer (LCAT), which decouples luminance and chrominance representations. Extensive experiments demonstrate that DSDNet outperforms state-of-the-art methods in both visual quality and quantitative evaluation, and achieves an inference speed $mathrm{ extbf{2.4x}}$ faster than the second-best method, highlighting its practical advantages. We provide an anonymous online demo at https://xxxxxxxxdsdnet.github.io/DSDNet/.

Problem

• Removes moiré artifacts from smartphone-captured screen images
• Addresses irreversible information loss in sRGB domain methods
• Improves efficiency over two-stage raw domain approaches

Innovation

• Single-stage raw domain demoiréing framework
• Synergic Attention with Dynamic Modulation module
• Luminance-Chrominance Adaptive Transformer for fidelity