Fantastic Breaks: A Dataset of Paired 3D Scans of Real-World Broken Objects and Their Complete Counterparts
Fantastic Breaks: A Dataset of Paired 3D Scans of Real-World Broken Objects and Their Complete Counterparts
复制标题
DOI：
10.1109/cvpr52729.2023.00454
复制
发表时间：
2023-03
期刊：
影响因子：
--
通讯作者：
N. Lamb;C. Palmer;Benjamin Molloy;Sean Banerjee;N. Banerjee
中科院分区：
文献类型：
--
作者：
N. Lamb;C. Palmer;Benjamin Molloy;Sean Banerjee;N. Banerjee
自动形状修复方法目前无法访问描述现实世界受损几何形状的数据集。我们展示了 Fantastic Breaks（以及在哪里找到它们：https://terascale-all-sensing-research-studio.github.io/FantasticBreaks），这是一个包含 150 个破碎物体的扫描、防水和清洁的 3D 网格的数据集，与完整的对应物配对并几何对齐。 Fantastic Breaks 包含类和材料标签、连接到损坏网格以生成完整网格的代理修复零件以及手动注释的断裂边界。通过对裂缝几何形状的详细分析，我们揭示了 Fantastic Breaks 与使用几何和物理方法生成的合成裂缝数据集之间的差异。我们展示了使用 Fantastic Breaks 进行的实验性形状修复评估，使用多种基于学习的方法，使用合成数据集进行预训练，并使用 Fantastic Breaks 子集进行重新训练。
Automated shape repair approaches currently lack access to datasets that describe real-world damaged geometry. We present Fantastic Breaks (and Where to Find Them: https://terascale-all-sensing-research-studio.github.io/FantasticBreaks), a dataset containing scanned, waterproofed, and cleaned 3D meshes for 150 broken objects, paired and geometrically aligned with complete counterparts. Fantastic Breaks contains class and material labels, proxy repair parts that join to broken meshes to generate complete meshes, and manually annotated fracture boundaries. Through a detailed analysis of fracture geometry, we reveal differences between Fantastic Breaks and synthetic fracture datasets generated using geometric and physics-based methods. We show experimental shape repair evaluation with Fantastic Breaks using multiple learning-based approaches pre-trained with synthetic datasets and re-trained with subset of Fantastic Breaks.