αMatte4K与μMatting：面向超微精度Alpha视频抠图的数据集与模型
$\alpha$Matte4K & $\mu$Matting: Dataset and Model for Ultra-Micro Precision Alpha Video Matting
提出分辨率无关的两阶段视频抠图框架，利用掩码自编码器定位粗哑光区域并用稀疏3D卷积精修，同时构建基于PBR的4K高质量数据集，为高分辨率视频抠图提供可复用的数据生成与模型设计思路。
原文依据 we introduce \muMatting, an innovative resolution-agnostic two-stage framework for video matting
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论高分辨率人体视频抠图旨在为半透明区域预测准确的alpha蒙版，同时确保帧间时序一致性。尽管已取得显著进展，现有方法仍无法在质量与效率之间实现令人满意的权衡，在主体稳定性、时序建模和计算成本方面存在局限。本文提出μMatting，一种创新的分辨率无关的两阶段视频抠图框架：（1）使用人像感知的掩码自编码器进行粗蒙版定位；（2）通过稀疏3D卷积对关键区域进行精修，并辅以时序调制器注入全局时空线索以增强一致性和上下文感知。从数据角度看，现有研究仍受限于数据集质量不足，包括（1）不完美标注导致的alpha分数值不准确，以及（2）任意前景-背景合成缺乏自然连贯性导致的视觉不一致。为解决此问题，我们推出αMatte4K，一个大规模4K分辨率人体视频抠图数据集，通过基于物理的渲染（PBR）实现准确标注和物理一致性。大量实验表明，μMatting在准确性和时空一致性上超越最优方法，同时αMatte4K提升了基线性能，推动了真实场景中的应用。项目已开源：https://github.com/kadatec/mu-Matting。
英文原文摘要 1525 字符
High-resolution human video matting aims to predict accurate alpha mattes for semi-transparent regions while ensuring temporal consistency across frames. Despite notable progress, current methods still fail to achieve a satisfactory trade-off between quality and efficiency, with limitations in subject stability, temporal modeling, and computational cost. In this paper, we introduce \muMatting, an innovative resolution-agnostic two-stage framework for video matting: (1) coarse matte localization using a portrait-aware masked autoencoder; (2) refinement of critical regions via sparse 3D convolution, augmented by a temporal modulator that injects global spatio-temporal cues for enhanced consistency and contextual awareness. From data perspective, existing research remains limited by the insufficient quality of datasets, including (1) inaccurate alpha fractional values resulting from imperfect annotation, and (2) visual inconsistencies arising from arbitrary foreground-background compositions that lack natural coherence. To address this, we introduce \alphaMatte4K, a large-scale 4K-resolution human video matting dataset, which achieves accurate annotations and physical consistency through physically based rendering (PBR). Extensive experiments show that \muMatting surpasses state-of-the-art methods in accuracy and spatio-temporal consistency, while \alphaMatte4K boosts baseline performance, driving applications in real-world scenarios. The project is open-sourced at https://github.com/kadatec/mu-Matting.