Search papers, labs, and topics across Lattice.
This paper introduces DCVC-MB, a neural video codec framework designed for efficient B-frame coding by leveraging an innovative IBP frame strategy and a spatio-temporal fusion model rooted in state-space methods. The framework employs an entropy-aware skipping mechanism to optimize coding times by selectively omitting certain latents, enhancing overall compression performance. Experimental results reveal that DCVC-MB achieves significant BD-rate reductions of up to 8.98% compared to existing neural video codecs and outperforms traditional codecs like VTM-19.0-LDP and VTM-19.0-RA by up to 30.45% and 1.81%, respectively.
Achieving up to 30.45% better compression than traditional codecs, DCVC-MB redefines efficiency in neural video coding.
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.
