6DGS：基于单张图像与3D Gaussian Splatting模型的6D位姿估计
6DGS: 6D Pose Estimation from a Single Image and a 3D Gaussian Splatting Model
AI总结：
提出6DGS方法，通过反转3D Gaussian Splatting渲染过程，从单张RGB图像和3DGS模型中以闭合形式估计6DoF相机位姿，无需初始化和迭代，在真实场景中旋转精度提升12%、平移精度提升22%，并接近实时运行。
摘要：
我们提出6DGS，用于在给定表示场景的3D Gaussian Splatting（3DGS）模型的情况下，估计目标RGB图像的相机位姿。6DGS避免了分析-综合方法（例如iNeRF）典型的迭代过程，这些方法也需要相机位姿的初始化才能收敛。相反，我们的方法通过反转3DGS渲染过程来估计6DoF位姿。从物体表面出发，我们定义了一个辐射Ellicell，它均匀地生成从参数化3DGS模型的每个椭球体发出的光线。每条Ellicell光线与每个椭球体的渲染参数相关联，这进而用于获得目标图像像素与投射光线之间的最佳绑定。然后对这些像素-光线绑定进行排序，以选择得分最高的光线束，它们的交点提供相机中心，进而得到相机旋转。所提出的解决方案免除了初始化所需的“先验”位姿，并以闭合形式求解6DoF位姿估计，无需迭代。此外，与现有的用于位姿估计的Novel View Synthesis（NVS）基线相比，6DGS在真实场景中可将整体平均旋转精度提高12%，平移精度提高22%，尽管不需要任何初始化位姿。同时，我们的方法接近实时运行，在消费级硬件上达到15fps。
英文摘要：
We propose 6DGS to estimate the camera pose of a target RGB image given a 3D Gaussian Splatting (3DGS) model representing the scene. 6DGS avoids the iterative process typical of analysis-by-synthesis methods (e.g. iNeRF) that also require an initialization of the camera pose in order to converge. Instead, our method estimates a 6DoF pose by inverting the 3DGS rendering process. Starting from the object surface, we define a radiant Ellicell that uniformly generates rays departing from each ellipsoid that parameterize the 3DGS model. Each Ellicell ray is associated with the rendering parameters of each ellipsoid, which in turn is used to obtain the best bindings between the target image pixels and the cast rays. These pixel-ray bindings are then ranked to select the best scoring bundle of rays, which their intersection provides the camera center and, in turn, the camera rotation. The proposed solution obviates the necessity of an "a priori" pose for initialization, and it solves 6DoF pose estimation in closed form, without the need for iterations. Moreover, compared to the existing Novel View Synthesis (NVS) baselines for pose estimation, 6DGS can improve the overall average rotational accuracy by 12% and translation accuracy by 22% on real scenes, despite not requiring any initialization pose. At the same time, our method operates near real-time, reaching 15fps on consumer hardware.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
