SemanticFeels: Semantic Labeling during In-Hand Manipulation

Anas Al Shikh Khalil, Haozhi Qi, Roberto Calandra · 2026

Read SemanticFeels: Semantic Labeling during In-Hand Manipulation by Anas Al Shikh Khalil, Haozhi Qi, Roberto Calandra (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-2602-14099.txt.

Description

Anas Al Shikh Khalil and two co-authors describe a 2026 arXiv system that lets robots sense what they hold. As robots join everyday tasks, they need to feel both shape and material while handling objects. The paper extends the NeuralFeels framework with semantic labels from vision and touch.

High-resolution Digit tactile readings feed a fine-tuned EfficientNet-B0 network that predicts local materials. Those predictions join an augmented signed distance field that predicts shape and material regions together. Tests show strong agreement between predicted and true materials on single- and multi-material objects. On one multi-material object, mean matching accuracy reached 79.87% across trials.

Catalog record

CREATORAnas Al Shikh Khalil, Haozhi Qi, Roberto Calandra
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE25 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602-14099

Browse the full library · About the Tower of Babel