PartNet
收藏资源简介:
PartNet是由斯坦福大学等机构创建的大型3D对象理解数据集,包含26,671个3D模型,涵盖24个对象类别,共计573,585个部分实例。数据集通过细粒度和层次化的部分信息标注,支持多种任务,如形状分析、动态3D场景建模和仿真、功能性分析等。PartNet的创建过程涉及专家定义的部分概念和层次化分割模板,以确保标注的一致性和完整性。该数据集主要应用于3D部分识别和分割任务,旨在推动机器对3D形状的细粒度和层次化理解。
PartNet is a large-scale 3D object understanding dataset developed by Stanford University and other institutions. It contains 26,671 3D models spanning 24 object categories, with a total of 573,585 part instances. Equipped with fine-grained and hierarchical part annotations, the dataset supports multiple tasks including shape analysis, dynamic 3D scene modeling and simulation, functional analysis, etc. The development of PartNet adopts expert-defined part concepts and hierarchical segmentation templates to guarantee the consistency and completeness of annotations. This dataset is primarily utilized for 3D part recognition and segmentation tasks, with the goal of advancing machines' fine-grained and hierarchical understanding of 3D shapes.




