Estimating Surface Normals with Depth Image Gradients for Fast and Accurate Registration

Yosuke Nakagawa, Hideaki Uchiyama, Hajime Nagahara, Rin-Ichiro Taniguchi

研究成果: 著書/レポートタイプへの貢献会議での発言

5 引用 (Scopus)

抄録

We present a fast registration framework with estimating surface normals from depth images. The key component in the framework is to utilize adjacent pixels and compute the normal at each pixel on a depth image by following three steps. First, image gradients on a depth image are computed with a 2D differential filtering. Next, two 3D gradient vectors are computed from horizontal and vertical depth image gradients. Finally, the normal vector is obtained from the cross product of the 3D gradient vectors. Since horizontal and vertical adjacent pixels at each pixel are considered composing a local 3D plane, the 3D gradient vectors are equivalent to tangent vectors of the plane. Compared with existing normal estimation based on fitting a plane to a point cloud, our depth image gradients based normal estimation is extremely faster because it needs only a few mathematical operations. We apply it to normal space sampling based 3D registration and validate the effectiveness of our registration framework by evaluating its accuracy and computational cost with a public dataset.

元の言語英語
ホスト出版物のタイトルProceedings - 2015 International Conference on 3D Vision, 3DV 2015
編集者Michael Brown, Jana Kosecka, Christian Theobalt
出版者Institute of Electrical and Electronics Engineers Inc.
ページ640-647
ページ数8
ISBN(電子版)9781467383325
DOI
出版物ステータス出版済み - 11 20 2015
イベント2015 International Conference on 3D Vision, 3DV 2015 - Lyon, フランス
継続期間: 10 19 201510 22 2015

その他

その他2015 International Conference on 3D Vision, 3DV 2015
フランス
Lyon
期間10/19/1510/22/15

Fingerprint

Pixels
Sampling
Costs

All Science Journal Classification (ASJC) codes

  • Computer Vision and Pattern Recognition

これを引用

Nakagawa, Y., Uchiyama, H., Nagahara, H., & Taniguchi, R-I. (2015). Estimating Surface Normals with Depth Image Gradients for Fast and Accurate Registration. : M. Brown, J. Kosecka, & C. Theobalt (版), Proceedings - 2015 International Conference on 3D Vision, 3DV 2015 (pp. 640-647). [7335535] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/3DV.2015.80

Estimating Surface Normals with Depth Image Gradients for Fast and Accurate Registration. / Nakagawa, Yosuke; Uchiyama, Hideaki; Nagahara, Hajime; Taniguchi, Rin-Ichiro.

Proceedings - 2015 International Conference on 3D Vision, 3DV 2015. 版 / Michael Brown; Jana Kosecka; Christian Theobalt. Institute of Electrical and Electronics Engineers Inc., 2015. p. 640-647 7335535.

研究成果: 著書/レポートタイプへの貢献会議での発言

Nakagawa, Y, Uchiyama, H, Nagahara, H & Taniguchi, R-I 2015, Estimating Surface Normals with Depth Image Gradients for Fast and Accurate Registration. : M Brown, J Kosecka & C Theobalt (版), Proceedings - 2015 International Conference on 3D Vision, 3DV 2015., 7335535, Institute of Electrical and Electronics Engineers Inc., pp. 640-647, 2015 International Conference on 3D Vision, 3DV 2015, Lyon, フランス, 10/19/15. https://doi.org/10.1109/3DV.2015.80
Nakagawa Y, Uchiyama H, Nagahara H, Taniguchi R-I. Estimating Surface Normals with Depth Image Gradients for Fast and Accurate Registration. : Brown M, Kosecka J, Theobalt C, 編集者, Proceedings - 2015 International Conference on 3D Vision, 3DV 2015. Institute of Electrical and Electronics Engineers Inc. 2015. p. 640-647. 7335535 https://doi.org/10.1109/3DV.2015.80
Nakagawa, Yosuke ; Uchiyama, Hideaki ; Nagahara, Hajime ; Taniguchi, Rin-Ichiro. / Estimating Surface Normals with Depth Image Gradients for Fast and Accurate Registration. Proceedings - 2015 International Conference on 3D Vision, 3DV 2015. 編集者 / Michael Brown ; Jana Kosecka ; Christian Theobalt. Institute of Electrical and Electronics Engineers Inc., 2015. pp. 640-647
@inproceedings{f7c89b4601234c4fb1477bd6e74ff67e,
title = "Estimating Surface Normals with Depth Image Gradients for Fast and Accurate Registration",
abstract = "We present a fast registration framework with estimating surface normals from depth images. The key component in the framework is to utilize adjacent pixels and compute the normal at each pixel on a depth image by following three steps. First, image gradients on a depth image are computed with a 2D differential filtering. Next, two 3D gradient vectors are computed from horizontal and vertical depth image gradients. Finally, the normal vector is obtained from the cross product of the 3D gradient vectors. Since horizontal and vertical adjacent pixels at each pixel are considered composing a local 3D plane, the 3D gradient vectors are equivalent to tangent vectors of the plane. Compared with existing normal estimation based on fitting a plane to a point cloud, our depth image gradients based normal estimation is extremely faster because it needs only a few mathematical operations. We apply it to normal space sampling based 3D registration and validate the effectiveness of our registration framework by evaluating its accuracy and computational cost with a public dataset.",
author = "Yosuke Nakagawa and Hideaki Uchiyama and Hajime Nagahara and Rin-Ichiro Taniguchi",
year = "2015",
month = "11",
day = "20",
doi = "10.1109/3DV.2015.80",
language = "English",
pages = "640--647",
editor = "Michael Brown and Jana Kosecka and Christian Theobalt",
booktitle = "Proceedings - 2015 International Conference on 3D Vision, 3DV 2015",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
address = "United States",

}

TY - GEN

T1 - Estimating Surface Normals with Depth Image Gradients for Fast and Accurate Registration

AU - Nakagawa, Yosuke

AU - Uchiyama, Hideaki

AU - Nagahara, Hajime

AU - Taniguchi, Rin-Ichiro

PY - 2015/11/20

Y1 - 2015/11/20

N2 - We present a fast registration framework with estimating surface normals from depth images. The key component in the framework is to utilize adjacent pixels and compute the normal at each pixel on a depth image by following three steps. First, image gradients on a depth image are computed with a 2D differential filtering. Next, two 3D gradient vectors are computed from horizontal and vertical depth image gradients. Finally, the normal vector is obtained from the cross product of the 3D gradient vectors. Since horizontal and vertical adjacent pixels at each pixel are considered composing a local 3D plane, the 3D gradient vectors are equivalent to tangent vectors of the plane. Compared with existing normal estimation based on fitting a plane to a point cloud, our depth image gradients based normal estimation is extremely faster because it needs only a few mathematical operations. We apply it to normal space sampling based 3D registration and validate the effectiveness of our registration framework by evaluating its accuracy and computational cost with a public dataset.

AB - We present a fast registration framework with estimating surface normals from depth images. The key component in the framework is to utilize adjacent pixels and compute the normal at each pixel on a depth image by following three steps. First, image gradients on a depth image are computed with a 2D differential filtering. Next, two 3D gradient vectors are computed from horizontal and vertical depth image gradients. Finally, the normal vector is obtained from the cross product of the 3D gradient vectors. Since horizontal and vertical adjacent pixels at each pixel are considered composing a local 3D plane, the 3D gradient vectors are equivalent to tangent vectors of the plane. Compared with existing normal estimation based on fitting a plane to a point cloud, our depth image gradients based normal estimation is extremely faster because it needs only a few mathematical operations. We apply it to normal space sampling based 3D registration and validate the effectiveness of our registration framework by evaluating its accuracy and computational cost with a public dataset.

UR - http://www.scopus.com/inward/record.url?scp=84961762758&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84961762758&partnerID=8YFLogxK

U2 - 10.1109/3DV.2015.80

DO - 10.1109/3DV.2015.80

M3 - Conference contribution

AN - SCOPUS:84961762758

SP - 640

EP - 647

BT - Proceedings - 2015 International Conference on 3D Vision, 3DV 2015

A2 - Brown, Michael

A2 - Kosecka, Jana

A2 - Theobalt, Christian

PB - Institute of Electrical and Electronics Engineers Inc.

ER -