WildDet3D Scaling Promptable 3D Detection in the Wild

2026

Read WildDet3D Scaling Promptable 3D Detection in the Wild (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-2604-08626.txt.

Description

Huang and seventeen colleagues build WildDet3D, a unified system for detecting 3D objects from a single image in the open world. It takes text, point, and box prompts and can use depth cues at inference time, unlike single prompt methods. Trained on WildDet3D-Data, over one million human verified images across 13,500 categories, it sets new state of the art marks on open world, Omni3D, and zero shot benchmarks. Depth cues add large gains on top.

Catalog record

YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE94372
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.08626

Browse the full library · About the Tower of Babel