Instance–Guided Unsupervised Domain Adaptation for Robotic Semantic Segmentation

Michele Antonazzi, Lorenzo Signorelli, Matteo Luperto, Nicola Basilico · 2026

Read Instance–Guided Unsupervised Domain Adaptation for Robotic Semantic Segmentation by Michele Antonazzi, Lorenzo Signorelli, Matteo Luperto, Nicola Basilico (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-2602-01389.txt.

Description

Michele Antonazzi and co-authors help robots adapt their sight on the job when new scenes differ from training scenes. Their plan starts from a volumetric 3D map and draws multi-view consistent pseudo-labels across the map. A foundation model then refines those labels with zero-shot instance segmentation, forcing instance-level coherence. The cleaned labels steer self-supervised fine-tuning at deployment time, with no ground-truth labels needed in the target domain. Real-world tests show steady gains over state-of-the-art multi-view consistency baselines.

Catalog record

CREATORMichele Antonazzi, Lorenzo Signorelli, Matteo Luperto, Nicola Basilico
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE43154
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602.01389

Browse the full library · About the Tower of Babel