🧪 Test?View on arXiv
ObjectTwin: Measuring the Asset Correspondence Gap for Capture-Anchored Digital Twins
Not provided in the abstract
digital twins3D reconstructionasset correspondenceRGB-D capture
2610.35904
Builder Relevance
1h ago70%
Abstract
This paper presents ObjectTwin, a measurement protocol for assessing the asset correspondence gap in digital twins, highlighting the importance of accurate object representation in captured scenes.
Reality Card
Core Claim
ObjectTwin establishes a measurable interface for asset correspondence, revealing that generated assets lose 0.302-0.497 MaskIoU when compared to scan-derived geometry.
Method / Result
Visible, anchored generated assets lose 0.302-0.497 MaskIoU under the replacement of object representation.
Limitations
The evaluated conservative pose refinement does not systematically recover the observed discrepancies.
Paper to code
Verified implementation resources so builders can test the paper’s claims instead of stopping at the abstract.
No verified implementation link has been attached yet. AIBuzzHub will keep this panel separate from unverified search results.