Hybrid Spatial-Temporal Entropy Modelling for Neural Video Compression

TL;DR

This paper introduces a novel entropy model for neural video compression that captures spatial-temporal dependencies and enables content-adaptive quantization, significantly improving compression efficiency over existing methods.

Contribution

It proposes a new entropy model that leverages spatial and temporal correlations and includes a content-adaptive quantization mechanism, advancing neural video codec performance.

Findings

Achieves 18.2% bitrate reduction on UVG dataset compared to H.266 (VTM).

Introduces a versatile entropy model for better rate-distortion performance.

Enables smooth rate adjustment through dynamic bit allocation.

Abstract

For neural video codec, it is critical, yet challenging, to design an efficient entropy model which can accurately predict the probability distribution of the quantized latent representation. However, most existing video codecs directly use the ready-made entropy model from image codec to encode the residual or motion, and do not fully leverage the spatial-temporal characteristics in video. To this end, this paper proposes a powerful entropy model which efficiently captures both spatial and temporal dependencies. In particular, we introduce the latent prior which exploits the correlation among the latent representation to squeeze the temporal redundancy. Meanwhile, the dual spatial prior is proposed to reduce the spatial redundancy in a parallel-friendly manner. In addition, our entropy model is also versatile. Besides estimating the probability distribution, our entropy model also generates the quantization step at spatial-channel-wise. This content-adaptive quantization mechanism not only helps our codec achieve the smooth rate adjustment in single model but also improves the final rate-distortion performance by dynamic bit allocation. Experimental results show that, powered by the proposed entropy model, our neural codec can achieve 18.2% bitrate saving on UVG dataset when compared with H.266 (VTM) using the highest compression ratio configuration. It makes a new milestone in the development of neural video codec. The codes are at https://github.com/microsoft/DCVC.