AmaraSpatial-10K: A Spatially and Semantically Aligned 3D Dataset for Spatial Computing and Embodied AI
Mohammad Sadegh Salehi, Alex Perkins, Igor Maurell, Ashkan Dabbagh, Raymond Wong · 2026
Read AmaraSpatial-10K: A Spatially and Semantically Aligned 3D Dataset for Spatial Computing and Embodied AI by Mohammad Sadegh Salehi, Alex Perkins, Igor Maurell, Ashkan Dabbagh, Raymond Wong (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-23018.txt.
Description
Salehi and four colleagues present AmaraSpatial-10K, a data set of over 10,000 synthetic 3D assets built for zero-shot use in robotics and spatial computing. Web-scale asset collections often suffer from wrong scaling, bad pivots, brittle shapes, and incomplete textures. Each asset ships metric-scaled and anchored, with split texture maps, a collision hull, a reference image, and text metadata. A reusable test suite shows strong retrieval and stability gains over matched subsets of existing collections.
Catalog record
| CREATOR | Mohammad Sadegh Salehi, Alex Perkins, Igor Maurell, Ashkan Dabbagh, Raymond Wong |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 68042 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.23018