通过可平移 RGB-A 分布学习器生成稳定透明度的视频
Video Generation with Stable Transparency via Shiftable RGB-A Distribution Learner
通过可平移的 RGB-A 分布学习，在 VAE 训练中引入透明度感知的双向扩散损失，并在噪声空间平移采样均值与高斯椭圆掩码，为透明视频生成提供稳定且可控的扩散采样策略。
原文依据 we propose a transparency-aware bidirectional diffusion loss during VAE training, which shifts the RGB-A distribution according to likelihood.
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论生成包含透明度 alpha 通道的 RGB-A 视频具有广泛的应用。然而，当前方法常因 RGB 与 alpha 之间的混淆而导致质量低下。本文通过学习可平移的 RGB-A 分布来解决这一问题。我们调整潜空间和噪声空间，在保持 RGB 分布的同时将 alpha 分布向外平移，从而在不损害 RGB 质量的情况下实现稳定的透明度生成。具体而言，对于潜空间，我们在变分自编码器训练中提出了一种透明度感知的双向扩散损失，该损失根据似然平移 RGB-A 分布。对于噪声空间，我们提出平移扩散噪声采样的均值，并应用高斯椭圆掩码以提供透明度引导和可控性。此外，我们构建了一个高质量的 RGB-A 视频数据集。与最优方法相比，我们的模型在视觉质量、自然度、透明度渲染、推理便利性和可控性方面均表现出色。发布的模型可在我们的网站获取：https://donghaotian123.github.io/Wan-Alpha/。
英文原文摘要 1140 字符
Generating RGB-A videos, which include alpha channels for transparency, has wide applications. However, current methods often suffer from low quality due to confusion between RGB and alpha. In this paper, we address this problem by learning shiftable RGB-A distributions. We adjust both the latent space and noise space, shifting the alpha distribution outward while preserving the RGB distribution, thereby enabling stable transparency generation without compromising RGB quality. Specifically, for the latent space, we propose a transparency-aware bidirectional diffusion loss during VAE training, which shifts the RGB-A distribution according to likelihood. For the noise space, we propose shifting the mean of diffusion noise sampling and applying a Gaussian ellipse mask to provide transparency guidance and controllability. Additionally, we construct a high-quality RGB-A video dataset. Compared to state-of-the-art methods, our model excels in visual quality, naturalness, transparency rendering, inference convenience, and controllability. The released model is available on our website: https://donghaotian123.github.io/Wan-Alpha/.