Neural radiance field using pixel-aligned features can render photo-realistic novel views. However, when pixel-aligned features are directly introduced to human avatar reconstruction, the rendering can only be conducted for still humans, rather than animatable avatars. In this paper, we propose AniPixel, a novel animatable and generalizable human avatar reconstruction method that leverages pixel-aligned features for body geometry prediction and RGB color blending. Technically, to align the canonical space with the target space and the observation space, we propose a bidirectional neural skinning field based on skeleton-driven deformation to establish the target-to-canonical and canonical-to-observation correspondences. Then, we disentangle the canonical body geometry into a normalized neutral-sized body and a subject-specific residual for better generalizability. As the geometry and appearance are closely related, we introduce pixel-aligned features to facilitate the body geometry prediction and detailed surface normals to reinforce the RGB color blending. Moreover, we devise a pose-dependent and view direction-related shading module to represent the local illumination variance. Experiments show that our AniPixel renders comparable novel views while delivering better novel pose animation results than state-of-the-art methods. The code will be released.

翻译：基于像素对齐特征的神经辐射场能够渲染出逼真的新视角图像。然而，当将像素对齐特征直接引入人体化身重建时，渲染仅能针对静态人体进行，无法实现可动画化的化身。本文提出AniPixel——一种新颖的可动画化且可泛化的人体化身重建方法，该方法利用像素对齐特征进行体几何预测与RGB颜色融合。在技术层面，为对齐规范空间、目标空间与观测空间，我们提出一种基于骨骼驱动形变的双向神经蒙皮场，以建立目标到规范空间与规范到观测空间的对应关系。接着，将规范体几何解耦为归一化的中性尺寸体与对象特定残差，以增强泛化能力。由于几何与外观紧密关联，我们引入像素对齐特征辅助体几何预测，并利用精细表面法向强化RGB颜色融合。此外，设计了一个依赖姿态与视角方向的着色模块，用于表征局部光照变化。实验表明，相较于现有最优方法，AniPixel在生成可比较的新视角图像的同时，能够呈现更优的新姿态动画结果。代码将开源。