Robust Vector Alignment Dataset
收藏资源简介:
该数据集名为'Robust Vector Alignment Dataset',由弗吉尼亚理工大学机械工程系的研究人员创建,旨在解决球形模式之间的旋转估计问题。数据集包含5个模板球形模式和每个模板700个源模式(共计3500个模式),这些模式通过100次随机旋转采样生成,涵盖了整个旋转空间,并具有不同级别的噪声和异常值。数据集适用于球形点模式配准、点云配准和球形图像配准等任务,有助于提升算法在现实复杂场景中的有效性和鲁棒性。
The dataset named 'Robust Vector Alignment Dataset' was created by researchers from the Department of Mechanical Engineering, Virginia Tech, aiming to address the rotation estimation problem between spherical patterns. The dataset includes 5 template spherical patterns and 700 source patterns for each template, totaling 3500 patterns. These patterns are generated via 100 rounds of random rotation sampling, covering the entire rotation space, and feature different levels of noise and outliers. This dataset is applicable to tasks including spherical point pattern registration, point cloud registration and spherical image registration, and helps enhance the effectiveness and robustness of algorithms in real-world complex scenarios.

- 1Correspondence-Free Fast and Robust Spherical Point Pattern Registration弗吉尼亚理工大学机械工程系 · 2025年



