非刚性结构从运动数据集

非刚性结构从运动数据集是由丹麦技术大学计算部创建的，旨在为非刚性结构从运动（nrsfm）的研究提供高质量的数据支持。该数据集包含60个不同的序列，涵盖了五种不同的物体/场景和变形，通过机械动画技术生成，确保了实验的可重复性。数据集还包括六种不同的摄像机运动，使用正交和全视角摄像机模型，以及通过自遮挡生成的真实结构缺失数据。该数据集的设计旨在提供一个严格的统计分析平台，以评估和推动nrsfm领域的技术发展。

The Non-rigid Structure from Motion (NRSFM) dataset was created by the Department of Computing, Technical University of Denmark, aiming to provide high-quality data support for research in the NRSFM field. This dataset contains 60 distinct sequences covering five different object/scene categories and deformation patterns, which are generated via mechanical animation techniques to ensure the reproducibility of experiments. Additionally, the dataset includes six distinct camera motions, utilizing both orthographic and perspective camera models, as well as realistic structural missing data generated through self-occlusion. The dataset is designed to serve as a rigorous statistical analysis platform for evaluating and advancing technological progress in the NRSFM domain.