Local shape feature fusion for improved matching, pose estimation and 3D object recognition

Research output: Contribution to journalJournal articleResearchpeer-review

168 Downloads (Pure)

Abstract

We provide new insights to the problem of shape feature description and matching, techniques that are often applied within 3D object recognition pipelines. We subject several state of the art features to systematic evaluations based on multiple datasets from different sources in a uniform manner. We have carefully prepared and performed a neutral test on the datasets for which the descriptors have shown good recognition performance. Our results expose an important fallacy of previous results, namely that the performance of the recognition system does not correlate well with the performance of the descriptor employed by the recognition system. In addition to this, we evaluate several aspects of the matching task, including the efficiency of the different features, and the potential in using dimension reduction. To arrive at better generalization properties, we introduce a method for fusing several feature matches with a limited processing overhead. Our fused feature matches provide a significant increase in matching accuracy, which is consistent over all tested datasets. Finally, we benchmark all features in a 3D object recognition setting, providing further evidence of the advantage of fused features, both in terms of accuracy and efficiency.
Original languageEnglish
Article number297
JournalSpringerPlus
Volume5
Issue number1
Number of pages33
ISSN2193-1801
DOIs
Publication statusPublished - 8. Mar 2016

Fingerprint

Benchmarking
Datasets

Keywords

  • 3D shape descriptors
  • 3D object recognition
  • Shape matching
  • Feature fusion

Cite this

@article{af15b23731244172acaa6508af89ee5e,
title = "Local shape feature fusion for improved matching, pose estimation and 3D object recognition",
abstract = "We provide new insights to the problem of shape feature description and matching, techniques that are often applied within 3D object recognition pipelines. We subject several state of the art features to systematic evaluations based on multiple datasets from different sources in a uniform manner. We have carefully prepared and performed a neutral test on the datasets for which the descriptors have shown good recognition performance. Our results expose an important fallacy of previous results, namely that the performance of the recognition system does not correlate well with the performance of the descriptor employed by the recognition system. In addition to this, we evaluate several aspects of the matching task, including the efficiency of the different features, and the potential in using dimension reduction. To arrive at better generalization properties, we introduce a method for fusing several feature matches with a limited processing overhead. Our fused feature matches provide a significant increase in matching accuracy, which is consistent over all tested datasets. Finally, we benchmark all features in a 3D object recognition setting, providing further evidence of the advantage of fused features, both in terms of accuracy and efficiency.",
keywords = "3D shape descriptors, 3D object recognition, Shape matching, Feature fusion",
author = "Buch, {Anders Glent} and Petersen, {Henrik Gordon} and Norbert Kr{\"u}ger",
year = "2016",
month = "3",
day = "8",
doi = "10.1186/s40064-016-1906-1",
language = "English",
volume = "5",
journal = "SpringerPlus",
issn = "2193-1801",
publisher = "Springer",
number = "1",

}

Local shape feature fusion for improved matching, pose estimation and 3D object recognition. / Buch, Anders Glent; Petersen, Henrik Gordon; Krüger, Norbert.

In: SpringerPlus, Vol. 5, No. 1, 297, 08.03.2016.

Research output: Contribution to journalJournal articleResearchpeer-review

TY - JOUR

T1 - Local shape feature fusion for improved matching, pose estimation and 3D object recognition

AU - Buch, Anders Glent

AU - Petersen, Henrik Gordon

AU - Krüger, Norbert

PY - 2016/3/8

Y1 - 2016/3/8

N2 - We provide new insights to the problem of shape feature description and matching, techniques that are often applied within 3D object recognition pipelines. We subject several state of the art features to systematic evaluations based on multiple datasets from different sources in a uniform manner. We have carefully prepared and performed a neutral test on the datasets for which the descriptors have shown good recognition performance. Our results expose an important fallacy of previous results, namely that the performance of the recognition system does not correlate well with the performance of the descriptor employed by the recognition system. In addition to this, we evaluate several aspects of the matching task, including the efficiency of the different features, and the potential in using dimension reduction. To arrive at better generalization properties, we introduce a method for fusing several feature matches with a limited processing overhead. Our fused feature matches provide a significant increase in matching accuracy, which is consistent over all tested datasets. Finally, we benchmark all features in a 3D object recognition setting, providing further evidence of the advantage of fused features, both in terms of accuracy and efficiency.

AB - We provide new insights to the problem of shape feature description and matching, techniques that are often applied within 3D object recognition pipelines. We subject several state of the art features to systematic evaluations based on multiple datasets from different sources in a uniform manner. We have carefully prepared and performed a neutral test on the datasets for which the descriptors have shown good recognition performance. Our results expose an important fallacy of previous results, namely that the performance of the recognition system does not correlate well with the performance of the descriptor employed by the recognition system. In addition to this, we evaluate several aspects of the matching task, including the efficiency of the different features, and the potential in using dimension reduction. To arrive at better generalization properties, we introduce a method for fusing several feature matches with a limited processing overhead. Our fused feature matches provide a significant increase in matching accuracy, which is consistent over all tested datasets. Finally, we benchmark all features in a 3D object recognition setting, providing further evidence of the advantage of fused features, both in terms of accuracy and efficiency.

KW - 3D shape descriptors

KW - 3D object recognition

KW - Shape matching

KW - Feature fusion

U2 - 10.1186/s40064-016-1906-1

DO - 10.1186/s40064-016-1906-1

M3 - Journal article

C2 - 27066334

VL - 5

JO - SpringerPlus

JF - SpringerPlus

SN - 2193-1801

IS - 1

M1 - 297

ER -