标题： SNeRV：视频的光谱保持神经表示 显示英文标题

标题： SNeRV: Spectra-preserving Neural Representation for Video

摘要： 神经视频表示（NeRV），采用神经网络对视频信号进行参数化，为视频表示引入了一种新方法。 然而，现有的基于NeRV的方法由于频谱偏差，在捕捉精细的空间细节和运动模式方面存在困难，其中神经网络学习高频（HF）成分的速度比低频（LF）成分慢。 在本文中，我们提出了频谱保持的NeRV（SNeRV），作为一种新方法，通过有效处理各种频率成分来增强隐式视频表示。 SNeRV使用二维离散小波变换（DWT）将视频分解为LF和HF特征，保留空间结构并直接解决频谱偏差问题。 为了平衡紧凑性，我们仅编码LF成分，而包含精细纹理的HF成分由解码器生成。 集成到主干中的专用模块，包括多分辨率融合单元（MFU）和高频恢复器（HFR），有助于表示。 此外，我们通过将扩展作为时间域的额外频率分解，将SNeRV扩展以有效捕捉相邻视频帧之间的时序相关性。 这种方法使我们能够将时空LF特征嵌入网络中，使用时间扩展的上采样块（TUBs）。 实验结果表明，SNeRV在捕捉精细细节方面优于现有的NeRV模型，并实现了增强的重建效果，使其成为隐式视频表示领域的一种有前景的方法。 代码可在 https://github.com/qwertja/SNeRV 获取。 显示英文摘要

摘要： Neural representation for video (NeRV), which employs a neural network to parameterize video signals, introduces a novel methodology in video representations. However, existing NeRV-based methods have difficulty in capturing fine spatial details and motion patterns due to spectral bias, in which a neural network learns high-frequency (HF) components at a slower rate than low-frequency (LF) components. In this paper, we propose spectra-preserving NeRV (SNeRV) as a novel approach to enhance implicit video representations by efficiently handling various frequency components. SNeRV uses 2D discrete wavelet transform (DWT) to decompose video into LF and HF features, preserving spatial structures and directly addressing the spectral bias issue. To balance the compactness, we encode only the LF components, while HF components that include fine textures are generated by a decoder. Specialized modules, including a multi-resolution fusion unit (MFU) and a high-frequency restorer (HFR), are integrated into a backbone to facilitate the representation. Furthermore, we extend SNeRV to effectively capture temporal correlations between adjacent video frames, by casting the extension as additional frequency decomposition to a temporal domain. This approach allows us to embed spatio-temporal LF features into the network, using temporally extended up-sampling blocks (TUBs). Experimental results demonstrate that SNeRV outperforms existing NeRV models in capturing fine details and achieves enhanced reconstruction, making it a promising approach in the field of implicit video representations. The codes are available at https://github.com/qwertja/SNeRV.