Frequency-aware Neural Representation for Videos

TL;DR

FaNeRV is a novel frequency-aware neural video representation that explicitly separates spectral components, leading to improved reconstruction quality and rate-distortion performance over existing INR-based methods.

Contribution

The paper introduces FaNeRV, a frequency-aware neural network with multi-resolution supervision and high-frequency injection, addressing spectral bias in INR-based video compression.

Findings

Outperforms state-of-the-art INR methods in experiments.

Achieves competitive rate-distortion performance with traditional codecs.

Effectively captures both global structures and fine textures.

Abstract

Implicit Neural Representations (INRs) have emerged as a promising paradigm for video compression. However, existing INR-based frameworks typically suffer from inherent spectral bias, which favors low-frequency components and leads to over-smoothed reconstructions and suboptimal rate-distortion performance. In this paper, we propose FaNeRV, a Frequency-aware Neural Representation for videos, which explicitly decouples low- and high-frequency components to enable efficient and faithful video reconstruction. FaNeRV introduces a multi-resolution supervision strategy that guides the network to progressively capture global structures and fine-grained textures through staged supervision . To further enhance high-frequency reconstruction, we propose a dynamic high-frequency injection mechanism that adaptively emphasizes challenging regions. In addition, we design a frequency-decomposed network module to improve feature modeling across different spectral bands. Extensive experiments on standard benchmarks demonstrate that FaNeRV significantly outperforms state-of-the-art INR methods and achieves competitive rate-distortion performance against traditional codecs.