ABO: Dataset and Benchmarks for Real-World 3D Object Understanding
This provides a resource for researchers in computer vision and graphics to improve 3D object understanding, but it is incremental as it focuses on dataset creation rather than novel algorithmic advances.
The authors tackled the problem of bridging real and virtual 3D worlds by introducing the ABO dataset, which includes product images, metadata, and 3D models for household objects, and they established benchmarks showing current state-of-the-art limits on tasks like 3D reconstruction and material estimation.
We introduce Amazon Berkeley Objects (ABO), a new large-scale dataset designed to help bridge the gap between real and virtual 3D worlds. ABO contains product catalog images, metadata, and artist-created 3D models with complex geometries and physically-based materials that correspond to real, household objects. We derive challenging benchmarks that exploit the unique properties of ABO and measure the current limits of the state-of-the-art on three open problems for real-world 3D object understanding: single-view 3D reconstruction, material estimation, and cross-domain multi-view object retrieval.