ABO

ABO是产品图像和逼真的，高分辨率的，基于物理的家用对象3D模型的数据集。我们介绍了Amazon Berkeley Objects (ABO)，这是一个新的大规模数据集，旨在帮助弥合真实和虚拟3D世界之间的鸿沟。ABO包含产品目录图像，元数据和艺术家创建的3D模型，这些模型具有复杂的几何形状和与真实的家用对象相对应的基于物理的材料。我们推导了具有挑战性的基准，这些基准利用了ABO的独特属性，并针对现实世界3D对象理解的三个开放问题测量了最新技术的当前极限: 单视图3D重建，材料估计和跨域多视图对象检索。

ABO is a dataset comprising product images, as well as realistic, high-resolution, physically based 3D models of household objects. We present Amazon Berkeley Objects (ABO), a novel large-scale dataset designed to bridge the gap between the real and virtual 3D worlds. ABO contains product catalog images, metadata, and artist-created 3D models with intricate geometries and physically based materials corresponding to real-world household objects. We develop challenging benchmarks that exploit the unique characteristics of ABO, and measure the current state-of-the-art performance limits for three open problems in real-world 3D object understanding: single-view 3D reconstruction, material estimation, and cross-domain multi-view object retrieval.