A range image refinement technique for multi-view 3D model reconstruction

S. Y. Park; Murali Subbarao

doi:10.1109/IM.2003.1240244

A range image refinement technique for multi-view 3D model reconstruction

S. Y. Park, Murali Subbarao

School of Electronics Engineering

Stony Brook University

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

2 Scopus citations

Abstract

We present a range image refinement technique for generating accurate 3D computer models of real objects. Range images obtained from a stereo-vision system typically experience geometric distortions on reconstructed 3D surfaces due to the inherent stereo matching problems such as occlusions or mismatchings. We introduce a range image refinement technique to correct such erroneous ranges by employing epipolar geometry of a multiview modelling system and the visual hull of an object. After registering multiple range images into a common coordinate system, we first determine if a 3D point in a range image is erroneous, by measuring registration of the point with its correspondences in other range images. The correspondences are determined on 3D contours which are inverse-projections of epipolar lines in other 2D silhouette images. Then the range of the point is refined onto the object's surface, if it is erroneous. We employ two techniques to search the correspondences fast. In case that there is no correspondence for an erroneous point, we refine the point onto the visual hull of the object. We show that refined range images yield better geometric structures in reconstructed 3D models.

Original language	English
Title of host publication	Proceedings - 4th International Conference on 3-D Digital Imaging and Modeling, 3DIM 2003
Publisher	IEEE Computer Society
Pages	147-154
Number of pages	8
ISBN (Electronic)	0769519911
DOIs	https://doi.org/10.1109/IM.2003.1240244
State	Published - 2003
Event	4th International Conference on 3-D Digital Imaging and Modeling, 3DIM 2003 - Banff, Canada Duration: 6 Oct 2003 → 10 Oct 2003

Publication series

Name	Proceedings of International Conference on 3-D Digital Imaging and Modeling, 3DIM
Volume	2003-January
ISSN (Print)	1550-6185

Conference

Conference	4th International Conference on 3-D Digital Imaging and Modeling, 3DIM 2003
Country/Territory	Canada
City	Banff
Period	6/10/03 → 10/10/03

Keywords

Computer vision
Electronic mail
Image reconstruction
Image sensors
Low pass filters
Nonlinear filters
Solid modeling
Stereo image processing
Stereo vision
Surface reconstruction

Access to Document

10.1109/IM.2003.1240244

Cite this

Park, S. Y., & Subbarao, M. (2003). A range image refinement technique for multi-view 3D model reconstruction. In Proceedings - 4th International Conference on 3-D Digital Imaging and Modeling, 3DIM 2003 (pp. 147-154). Article 1240244 (Proceedings of International Conference on 3-D Digital Imaging and Modeling, 3DIM; Vol. 2003-January). IEEE Computer Society. https://doi.org/10.1109/IM.2003.1240244

@inproceedings{265628bb14f441dc844e821d8748c259,

title = "A range image refinement technique for multi-view 3D model reconstruction",

abstract = "We present a range image refinement technique for generating accurate 3D computer models of real objects. Range images obtained from a stereo-vision system typically experience geometric distortions on reconstructed 3D surfaces due to the inherent stereo matching problems such as occlusions or mismatchings. We introduce a range image refinement technique to correct such erroneous ranges by employing epipolar geometry of a multiview modelling system and the visual hull of an object. After registering multiple range images into a common coordinate system, we first determine if a 3D point in a range image is erroneous, by measuring registration of the point with its correspondences in other range images. The correspondences are determined on 3D contours which are inverse-projections of epipolar lines in other 2D silhouette images. Then the range of the point is refined onto the object's surface, if it is erroneous. We employ two techniques to search the correspondences fast. In case that there is no correspondence for an erroneous point, we refine the point onto the visual hull of the object. We show that refined range images yield better geometric structures in reconstructed 3D models.",

keywords = "Computer vision, Electronic mail, Image reconstruction, Image sensors, Low pass filters, Nonlinear filters, Solid modeling, Stereo image processing, Stereo vision, Surface reconstruction",

author = "Park, {S. Y.} and Murali Subbarao",

note = "Publisher Copyright: {\textcopyright} 2003 IEEE.; 4th International Conference on 3-D Digital Imaging and Modeling, 3DIM 2003 ; Conference date: 06-10-2003 Through 10-10-2003",

year = "2003",

doi = "10.1109/IM.2003.1240244",

language = "English",

series = "Proceedings of International Conference on 3-D Digital Imaging and Modeling, 3DIM",

publisher = "IEEE Computer Society",

pages = "147--154",

booktitle = "Proceedings - 4th International Conference on 3-D Digital Imaging and Modeling, 3DIM 2003",

address = "United States",

}

Park, SY & Subbarao, M 2003, A range image refinement technique for multi-view 3D model reconstruction. in Proceedings - 4th International Conference on 3-D Digital Imaging and Modeling, 3DIM 2003., 1240244, Proceedings of International Conference on 3-D Digital Imaging and Modeling, 3DIM, vol. 2003-January, IEEE Computer Society, pp. 147-154, 4th International Conference on 3-D Digital Imaging and Modeling, 3DIM 2003, Banff, Canada, 6/10/03. https://doi.org/10.1109/IM.2003.1240244

A range image refinement technique for multi-view 3D model reconstruction. / Park, S. Y.; Subbarao, Murali.
Proceedings - 4th International Conference on 3-D Digital Imaging and Modeling, 3DIM 2003. IEEE Computer Society, 2003. p. 147-154 1240244 (Proceedings of International Conference on 3-D Digital Imaging and Modeling, 3DIM; Vol. 2003-January).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - A range image refinement technique for multi-view 3D model reconstruction

AU - Park, S. Y.

AU - Subbarao, Murali

PY - 2003

Y1 - 2003

N2 - We present a range image refinement technique for generating accurate 3D computer models of real objects. Range images obtained from a stereo-vision system typically experience geometric distortions on reconstructed 3D surfaces due to the inherent stereo matching problems such as occlusions or mismatchings. We introduce a range image refinement technique to correct such erroneous ranges by employing epipolar geometry of a multiview modelling system and the visual hull of an object. After registering multiple range images into a common coordinate system, we first determine if a 3D point in a range image is erroneous, by measuring registration of the point with its correspondences in other range images. The correspondences are determined on 3D contours which are inverse-projections of epipolar lines in other 2D silhouette images. Then the range of the point is refined onto the object's surface, if it is erroneous. We employ two techniques to search the correspondences fast. In case that there is no correspondence for an erroneous point, we refine the point onto the visual hull of the object. We show that refined range images yield better geometric structures in reconstructed 3D models.

AB - We present a range image refinement technique for generating accurate 3D computer models of real objects. Range images obtained from a stereo-vision system typically experience geometric distortions on reconstructed 3D surfaces due to the inherent stereo matching problems such as occlusions or mismatchings. We introduce a range image refinement technique to correct such erroneous ranges by employing epipolar geometry of a multiview modelling system and the visual hull of an object. After registering multiple range images into a common coordinate system, we first determine if a 3D point in a range image is erroneous, by measuring registration of the point with its correspondences in other range images. The correspondences are determined on 3D contours which are inverse-projections of epipolar lines in other 2D silhouette images. Then the range of the point is refined onto the object's surface, if it is erroneous. We employ two techniques to search the correspondences fast. In case that there is no correspondence for an erroneous point, we refine the point onto the visual hull of the object. We show that refined range images yield better geometric structures in reconstructed 3D models.

KW - Computer vision

KW - Electronic mail

KW - Image reconstruction

KW - Image sensors

KW - Low pass filters

KW - Nonlinear filters

KW - Solid modeling

KW - Stereo image processing

KW - Stereo vision

KW - Surface reconstruction

UR - http://www.scopus.com/inward/record.url?scp=21944441270&partnerID=8YFLogxK

U2 - 10.1109/IM.2003.1240244

DO - 10.1109/IM.2003.1240244

M3 - Conference contribution

AN - SCOPUS:21944441270

T3 - Proceedings of International Conference on 3-D Digital Imaging and Modeling, 3DIM

SP - 147

EP - 154

BT - Proceedings - 4th International Conference on 3-D Digital Imaging and Modeling, 3DIM 2003

PB - IEEE Computer Society

T2 - 4th International Conference on 3-D Digital Imaging and Modeling, 3DIM 2003

Y2 - 6 October 2003 through 10 October 2003

ER -

A range image refinement technique for multi-view 3D model reconstruction

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this