DCVC-MB: Neural B-Frame Video Compression using State Space Models

Researchers propose DCVC-Mamba (DCVC-MB), a neural video codec for B-frame coding that uses state-space models for spatio-temporal fusion and an entropy-aware skipping mechanism. The method achieves BD-rate reductions of up to 8.98% over prior neural codecs and up to 30.45% over the VTM-19.0-LDP benchmark, advancing neural video compression.

arXiv:2607.14305v1 Announce Type: new Abstract: In this paper we propose DCVC-Mamba DCVC-MB , a neural video codec framework for B-frame coding. Our approach incorporates an IBP frame strategy for low-delay B-frame coding, a spatio-temporal fusion model based on state-space models for bidirectional temporal prediction, and an entropy-aware skipping mechanism that selectively omits coding certain latents to reduce entropy coding times. In addition to our model contributions we also implement two inference-time strategies that enhance compression performance. Experimental evaluation shows that DCVC-MB compares favorably to existing NVCs and traditional codecs. The method demonstrates BD-rate reductions of up to $8.98\%$ on average compared to prior neural video codecs, and improvements of up to $30.45\%$ and $1.81\%$ over the VTM-19.0-LDP and VTM-19.0-RA Inter-GoP=16 benchmarks, respectively, contributing to advances in neural video compression.