arXivDaily arXiv每日学术速递 周一至周五更新
​通知：联邦节假日调休：哥伦布日 / 原住民日（Columbus Day / Indigenous Peoples' Day），今日（13日）无更新。
arXiv 2510.26117cs.CV

姿态估计与3D高斯点的联合优化

JOGS: Joint Optimization of Pose Estimation and 3D Gaussian Splatting

Key Laboratory of Big Data & Artificial Intelligence in Transportation (Beijing Jiaotong University), Ministry of Education, China（大数据与人工智能在交通中的关键实验室（北京交通大学），教育部，中国）；School of Computer Science and Technology, Beijing Jiaotong University（计算机科学与技术学院，北京交通大学）；Department of Artificial Intelligence, Westlake University（人工智能系，西湖大学）

Xianben Yang, Yuxuan Li, Tao Wang, Tao Wang, Yi Jin, Yidong Li, Haibin Ling

阅读 PDF
arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文提出了一种联合优化姿态估计与3D高斯点的方法，通过可微渲染和定制光流算法提升重建质量和姿态准确性。

摘要：

传统的新视角合成方法严重依赖于外部相机姿态估计工具，如COLMAP，这往往引入计算瓶颈并传播误差。为了解决这些挑战，我们提出了一种统一的框架，通过联合优化3D高斯点和相机姿态，而无需预校准输入。我们的方法通过一种新颖的联合优化策略，迭代地优化3D高斯参数并更新相机姿态，从而同时提高场景重建保真度和姿态估计准确性。关键创新在于将联合优化分解为两个交替阶段：首先，通过可微渲染更新3D高斯参数，姿态固定；其次，利用定制的3D光流算法优化相机姿态，该算法结合了几何和光度约束。这种形式逐步减少投影误差，特别是在视角变化大和特征分布稀疏的挑战性场景中，传统方法往往力不从心。在多个数据集上的广泛评估表明，我们的方法在重建质量上显著优于现有的COLMAP-free技术，并且在一般情况下也超过了标准的COLMAP基线方法。

英文摘要：

Traditional novel view synthesis methods heavily rely on external camera pose estimation tools such as COLMAP, which often introduce computational bottlenecks and propagate errors. To address these challenges, we propose a unified framework that jointly optimizes 3D Gaussian points and camera poses without requiring pre-calibrated inputs. Our approach iteratively refines 3D Gaussian parameters and updates camera poses through a novel co-optimization strategy, ensuring simultaneous improvements in scene reconstruction fidelity and pose estimation accuracy. The key innovation lies in decoupling the joint optimization into two interleaved phases: first, updating 3D Gaussian parameters via differentiable rendering with fixed poses, and second, refining camera poses using a customized 3D optical flow algorithm that incorporates geometric and photometric constraints. This formulation progressively reduces projection errors, particularly in challenging scenarios with large viewpoint variations and sparse feature distributions, where traditional methods struggle. Extensive evaluations on multiple datasets demonstrate that our approach significantly outperforms existing COLMAP-free techniques in reconstruction quality, and also surpasses the standard COLMAP-based baseline in general.

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。

↑