SRNeRV: A Scale-wise Recursive Framework for Neural Video Representation

TL;DR

SRNeRV introduces a parameter-efficient recursive framework for neural video representation that leverages scale self-similarity, reducing redundancy and boosting rate-distortion performance in implicit neural representations.

Contribution

It proposes a novel shared architecture with a hybrid sharing scheme, replacing multi-scale stacking with recursive parameter sharing to improve efficiency and performance.

Findings

Significant reduction in model size due to shared modules.

Improved rate-distortion performance in INR scenarios.

Effective learning of scale-specific spatial patterns.

Abstract

Implicit Neural Representations (INRs) have emerged as a promising paradigm for video representation and compression. However, existing multi-scale INR generators often suffer from significant parameter redundancy by stacking independent processing blocks for each scale. Inspired by the principle of scale self-similarity in the generation process, we propose SRNeRV, a novel scale-wise recursive framework that replaces this stacked design with a parameter-efficient shared architecture. The core of our approach is a hybrid sharing scheme derived from decoupling the processing block into a scale-specific spatial mixing module and a scale-invariant channel mixing module. We recursively apply the same shared channel mixing module, which contains the majority of the parameters, across all scales, significantly reducing the model size while preserving the crucial capacity to learn scale-specific spatial patterns. Extensive experiments demonstrate that SRNeRV achieves a significant rate-distortion performance boost, especially in INR-friendly scenarios, validating that our sharing scheme successfully amplifies the core strengths of the INR paradigm.