Human-Interpretable Uncertainty Explanations for Point Cloud Registration

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Gaus, Johannes A., Schneider, Loris, Shi, Yitian, Lee, Jongseok, Rayyes, Rania, Triebel, Rudolph
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909803819827200
author Gaus, Johannes A.
Schneider, Loris
Shi, Yitian
Lee, Jongseok
Rayyes, Rania
Triebel, Rudolph
author_facet Gaus, Johannes A.
Schneider, Loris
Shi, Yitian
Lee, Jongseok
Rayyes, Rania
Triebel, Rudolph
contents In this paper, we address the point cloud registration problem, where well-known methods like ICP fail under uncertainty arising from sensor noise, pose-estimation errors, and partial overlap due to occlusion. We develop a novel approach, Gaussian Process Concept Attribution (GP-CA), which not only quantifies registration uncertainty but also explains it by attributing uncertainty to well-known sources of errors in registration problems. Our approach leverages active learning to discover new uncertainty sources in the wild by querying informative instances. We validate GP-CA on three publicly available datasets and in our real-world robot experiment. Extensive ablations substantiate our design choices. Our approach outperforms other state-of-the-art methods in terms of runtime, high sample-efficiency with active learning, and high accuracy. Our real-world experiment clearly demonstrates its applicability. Our video also demonstrates that GP-CA enables effective failure-recovery behaviors, yielding more robust robotic perception.
format Preprint
id arxiv_https___arxiv_org_abs_2509_18786
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Human-Interpretable Uncertainty Explanations for Point Cloud Registration
Gaus, Johannes A.
Schneider, Loris
Shi, Yitian
Lee, Jongseok
Rayyes, Rania
Triebel, Rudolph
Robotics
Computer Vision and Pattern Recognition
In this paper, we address the point cloud registration problem, where well-known methods like ICP fail under uncertainty arising from sensor noise, pose-estimation errors, and partial overlap due to occlusion. We develop a novel approach, Gaussian Process Concept Attribution (GP-CA), which not only quantifies registration uncertainty but also explains it by attributing uncertainty to well-known sources of errors in registration problems. Our approach leverages active learning to discover new uncertainty sources in the wild by querying informative instances. We validate GP-CA on three publicly available datasets and in our real-world robot experiment. Extensive ablations substantiate our design choices. Our approach outperforms other state-of-the-art methods in terms of runtime, high sample-efficiency with active learning, and high accuracy. Our real-world experiment clearly demonstrates its applicability. Our video also demonstrates that GP-CA enables effective failure-recovery behaviors, yielding more robust robotic perception.
title Human-Interpretable Uncertainty Explanations for Point Cloud Registration
topic Robotics
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2509.18786