A Novel 2.5D Feature Descriptor Compensating for Depth Rotation

Publikation: Bidrag til bog/antologi/rapport/konference-proceedingKonferencebidrag i proceedingsForskningpeer review

Resumé

We introduce a novel type of local image descriptor based on Gabor filter responses. Our method operates on RGB-D images. We use the depth information to compensate for perspective distortions caused by out-of-plane rotations. The descriptor contains the responses of a multi-resolution Gabor bank. Contrary to existing methods that rely on a dominant orientation estimate to achieve rotation invariance, we utilize the orientation information in the Gabor bank to achieve rotation invariance during the matching stage. Compared to SIFT and a recent also projective distortion compensating descriptor proposed for RGB-D data, our method achieves a significant increase in accuracy when tested on a wide-baseline RGB-D matching dataset.
OriginalsprogEngelsk
TitelProceedings of the 12th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications
Vol/bind4
ForlagSCITEPRESS Digital Library
Publikationsdato2017
Sider158-166
ISBN (Elektronisk)978-989-758-225-7
DOI
StatusUdgivet - 2017
Begivenhed12th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Porto, Portugal
Varighed: 27. feb. 20171. mar. 2017
Konferencens nummer: 12

Konference

Konference12th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications
Nummer12
LandPortugal
ByPorto
Periode27/02/201701/03/2017

Fingeraftryk

Invariance
Gabor filters

Citer dette

Hagelskjær, F., Buch, A. G., & Krüger, N. (2017). A Novel 2.5D Feature Descriptor Compensating for Depth Rotation. I Proceedings of the 12th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (Bind 4, s. 158-166). SCITEPRESS Digital Library. https://doi.org/10.5220/0006123201580166
Hagelskjær, Frederik ; Buch, Anders Glent ; Krüger, Norbert. / A Novel 2.5D Feature Descriptor Compensating for Depth Rotation. Proceedings of the 12th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications. Bind 4 SCITEPRESS Digital Library, 2017. s. 158-166
@inproceedings{62fa1b08320c44e1ac490d0795036a0a,
title = "A Novel 2.5D Feature Descriptor Compensating for Depth Rotation",
abstract = "We introduce a novel type of local image descriptor based on Gabor filter responses. Our method operates on RGB-D images. We use the depth information to compensate for perspective distortions caused by out-of-plane rotations. The descriptor contains the responses of a multi-resolution Gabor bank. Contrary to existing methods that rely on a dominant orientation estimate to achieve rotation invariance, we utilize the orientation information in the Gabor bank to achieve rotation invariance during the matching stage. Compared to SIFT and a recent also projective distortion compensating descriptor proposed for RGB-D data, our method achieves a significant increase in accuracy when tested on a wide-baseline RGB-D matching dataset.",
keywords = "Gabor filter, 2.5D, Depth rotation compensation, Descriptor",
author = "Frederik Hagelskj{\ae}r and Buch, {Anders Glent} and Norbert Kr{\"u}ger",
year = "2017",
doi = "10.5220/0006123201580166",
language = "English",
volume = "4",
pages = "158--166",
booktitle = "Proceedings of the 12th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications",
publisher = "SCITEPRESS Digital Library",

}

Hagelskjær, F, Buch, AG & Krüger, N 2017, A Novel 2.5D Feature Descriptor Compensating for Depth Rotation. i Proceedings of the 12th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications. bind 4, SCITEPRESS Digital Library, s. 158-166, 12th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, Porto, Portugal, 27/02/2017. https://doi.org/10.5220/0006123201580166

A Novel 2.5D Feature Descriptor Compensating for Depth Rotation. / Hagelskjær, Frederik; Buch, Anders Glent; Krüger, Norbert.

Proceedings of the 12th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications. Bind 4 SCITEPRESS Digital Library, 2017. s. 158-166.

Publikation: Bidrag til bog/antologi/rapport/konference-proceedingKonferencebidrag i proceedingsForskningpeer review

TY - GEN

T1 - A Novel 2.5D Feature Descriptor Compensating for Depth Rotation

AU - Hagelskjær, Frederik

AU - Buch, Anders Glent

AU - Krüger, Norbert

PY - 2017

Y1 - 2017

N2 - We introduce a novel type of local image descriptor based on Gabor filter responses. Our method operates on RGB-D images. We use the depth information to compensate for perspective distortions caused by out-of-plane rotations. The descriptor contains the responses of a multi-resolution Gabor bank. Contrary to existing methods that rely on a dominant orientation estimate to achieve rotation invariance, we utilize the orientation information in the Gabor bank to achieve rotation invariance during the matching stage. Compared to SIFT and a recent also projective distortion compensating descriptor proposed for RGB-D data, our method achieves a significant increase in accuracy when tested on a wide-baseline RGB-D matching dataset.

AB - We introduce a novel type of local image descriptor based on Gabor filter responses. Our method operates on RGB-D images. We use the depth information to compensate for perspective distortions caused by out-of-plane rotations. The descriptor contains the responses of a multi-resolution Gabor bank. Contrary to existing methods that rely on a dominant orientation estimate to achieve rotation invariance, we utilize the orientation information in the Gabor bank to achieve rotation invariance during the matching stage. Compared to SIFT and a recent also projective distortion compensating descriptor proposed for RGB-D data, our method achieves a significant increase in accuracy when tested on a wide-baseline RGB-D matching dataset.

KW - Gabor filter

KW - 2.5D

KW - Depth rotation compensation

KW - Descriptor

U2 - 10.5220/0006123201580166

DO - 10.5220/0006123201580166

M3 - Article in proceedings

VL - 4

SP - 158

EP - 166

BT - Proceedings of the 12th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications

PB - SCITEPRESS Digital Library

ER -

Hagelskjær F, Buch AG, Krüger N. A Novel 2.5D Feature Descriptor Compensating for Depth Rotation. I Proceedings of the 12th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications. Bind 4. SCITEPRESS Digital Library. 2017. s. 158-166 https://doi.org/10.5220/0006123201580166