LLVD：面向盲视频去噪的基于 LSTM 的潜在空间显式运动建模
LLVD: LSTM-based Explicit Motion Modeling in Latent Space for Blind Video Denoising
AI总结：
本文提出端到端盲视频去噪模型 LLVD，通过在编码潜在特征域结合时空特征提取与 LSTM 显式运动建模，在降低复杂度的同时提升 RAW 去噪性能并减少闪烁。
摘要：
视频恢复在修复退化视频内容方面发挥着关键作用，其通过纠正采集（传感器噪声、运动模糊等）、保存/共享（压缩、缩放等）以及编辑过程中引入的各类退化所造成的缺陷来实现。本文介绍了一种专为视频采集期间引入噪声的场景设计的新算法，旨在通过减少不需要的噪声伪影来提升视频的视觉质量。我们提出了潜在空间 LSTM 视频去噪器（Latent space LSTM Video Denoiser，LLVD），这是一种端到端的盲去噪模型。LLVD 独特地结合了空间和时间特征提取，在编码特征域内采用长短期记忆网络（Long Short Term Memory，LSTM）。这种 LSTM 层的集成对于保持恢复视频的连续性并最大限度减少闪烁至关重要。此外，在编码特征域中处理帧显著减少了计算量，从而形成非常轻量级的架构。LLVD 的盲去噪特性使其能够灵活适用于真实、野外的去噪场景，在这些场景中无法获得关于噪声特性的先验信息。实验表明，LLVD 在合成噪声和采集噪声方面均表现出优异性能。具体而言，LLVD 在 RAW 去噪任务上比当前最先进水平（State-Of-The-Art，SOTA）高出 0.3dB，同时计算复杂度降低了 59%。
英文摘要：
Video restoration plays a pivotal role in revitalizing degraded video content by rectifying imperfections caused by various degradations introduced during capturing (sensor noise, motion blur, etc.), saving/sharing (compression, resizing, etc.) and editing. This paper introduces a novel algorithm designed for scenarios where noise is introduced during video capture, aiming to enhance the visual quality of videos by reducing unwanted noise artifacts. We propose the Latent space LSTM Video Denoiser (LLVD), an end-to-end blind denoising model. LLVD uniquely combines spatial and temporal feature extraction, employing Long Short Term Memory (LSTM) within the encoded feature domain. This integration of LSTM layers is crucial for maintaining continuity and minimizing flicker in the restored video. Moreover, processing frames in the encoded feature domain significantly reduces computations, resulting in a very lightweight architecture. LLVD's blind nature makes it versatile for real, in-the-wild denoising scenarios where prior information about noise characteristics is not available. Experiments reveal that LLVD demonstrates excellent performance for both synthetic and captured noise. Specifically, LLVD surpasses the current State-Of-The-Art (SOTA) in RAW denoising by 0.3dB, while also achieving a 59\% reduction in computational complexity.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
