高分辨率神经视频编码：双向置信度引导的参考信息建模
High Resolution Neural Video Coding with Bi-directional Confidence-Guided Reference Information Modeling
针对高分辨率视频中光流不可靠和双向预测融合失真的问题，提出置信度引导的非均匀参考信息融合与多尺度运动表示，可复用于视频压缩、帧插值等需要双向时序建模的任务。
原文依据 we present the first High-Resolution bi-directional neural video coding method, termed as HR-NVC
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论利用双向上下文预测长期以来被认为是提升神经视频编码压缩效率的关键方向。然而，现有的神经B帧编解码器性能增益仍然有限，尤其是在运动幅度大的高分辨率视频中，光流估计变得不可靠，而均衡的预测融合会引入失真。为解决这些挑战，我们提出了首个高分辨率双向神经视频编码方法，称为HR-NVC，它非均匀地整合来自两个时序方向的置信度引导预测线索，以实现更可靠、更高效的压缩。具体而言，我们提出了时空锚定运动估计，引入虚拟锚框帧和低分辨率先验，显著提升了大位移下的估计鲁棒性。我们进一步设计了分层运动表示，将多尺度运动与时序参考相融合，实现了跨分辨率运动可靠性的紧凑自适应建模。最后，双向上下文非对称协调模块执行置信度引导的双向参考融合，有效抑制不可靠上下文，并在遮挡和场景转换区域附近恢复结构一致性。值得注意的是，我们的模型是首个在4K分辨率视频上评估的端到端优化视频编解码器，为更高分辨率神经视频编码建立了新的基准，并在神经B帧编解码器中达到了最优性能。
英文原文摘要 1543 字符
Exploiting bi-directional context prediction has long been recognized as a key direction for improving compression efficiency in neural video coding. However, existing neural B-frame codecs still exhibit limited performance gains, particularly in high-resolution videos with large motion, where optical flow estimation becomes unreliable and balanced prediction fusion introduces distortions. To address these challenges, we present the first High-Resolution bi-directional neural video coding method, termed as HR-NVC, which non-uniformly integrates confidence-guided predictive cues from both temporal directions to achieve more reliable and efficient compression. Specifically, we propose Spatio-Temporal Anchored Motion Estimation, which introduces virtual anchor frames and low-resolution priors to significantly improve estimation robustness under large displacements. We further design a Hierarchical Motion Representation that converges multi-scale motion with temporal references, enabling compact and adaptive modeling of motion reliability across resolutions. Finally, a Bi-Contextual Asymmetric Harmonization module performs confidence-guided fusion of bidirectional references, effectively suppressing unreliable contexts and restoring structural consistency near occlusion and scene transition regions. Notably, our model is the first end-to-end-optimized video codec evaluated on 4K-resolution videos, establishing a new benchmark for higher-resolution NVC and achieving state-of-the-art performance among neural B-frame codecs.