Qualitative structure from motion

Daphna Weinshall

doi:10.5555/118850.118898

Qualitative structure from motion

Daphna Weinshall^*

^*Corresponding author for this work

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

Abstract

Exact structure from motion is an ill-posed computation and therefore very sensitive to noise. In this work I describe how a qualitative shape representation, based on the sign of the Gaussian curvature, can be computed directly from motion disparities, without the computation of an exact depth map or the directions of surface normals. I show that humans can judge the curvature sense of three points undergoing 3D motion from two, three and four views with success rate significantly above chance. A simple RBF net has been trained to perform the same task.

Original language	English
Title of host publication	Advances in Neural Information Processing Systems 3, NIPS 1990
Editors	Richard Lippmann, John E. Moody, David S. Touretzky
Publisher	Neural information processing systems foundation
Pages	356-362
Number of pages	7
ISBN (Electronic)	1558601848, 9781558601840
DOIs	https://doi.org/10.5555/118850.118898
State	Published - 1990
Externally published	Yes
Event	3rd Advances in Neural Information Processing Systems, NIPS 1990 - Denver, United States Duration: 26 Nov 1990 → 29 Nov 1990

Publication series

Name	Advances in Neural Information Processing Systems
Volume	3
ISSN (Print)	1049-5258

Conference

Conference	3rd Advances in Neural Information Processing Systems, NIPS 1990
Country/Territory	United States
City	Denver
Period	26/11/90 → 29/11/90

Bibliographical note

Publisher Copyright:
© 1990 Neural information processing systems foundation. All rights reserved.

Access to Document

10.5555/118850.118898

Cite this

@inproceedings{27c58ec6df24414baadb993d876fd8af,

title = "Qualitative structure from motion",

abstract = "Exact structure from motion is an ill-posed computation and therefore very sensitive to noise. In this work I describe how a qualitative shape representation, based on the sign of the Gaussian curvature, can be computed directly from motion disparities, without the computation of an exact depth map or the directions of surface normals. I show that humans can judge the curvature sense of three points undergoing 3D motion from two, three and four views with success rate significantly above chance. A simple RBF net has been trained to perform the same task.",

author = "Daphna Weinshall",

note = "Publisher Copyright: {\textcopyright} 1990 Neural information processing systems foundation. All rights reserved.; 3rd Advances in Neural Information Processing Systems, NIPS 1990 ; Conference date: 26-11-1990 Through 29-11-1990",

year = "1990",

doi = "10.5555/118850.118898",

language = "אנגלית",

series = "Advances in Neural Information Processing Systems",

publisher = "Neural information processing systems foundation",

pages = "356--362",

editor = "Richard Lippmann and Moody, \{John E.\} and Touretzky, \{David S.\}",

booktitle = "Advances in Neural Information Processing Systems 3, NIPS 1990",

}

Weinshall, D 1990, Qualitative structure from motion. in R Lippmann, JE Moody & DS Touretzky (eds), Advances in Neural Information Processing Systems 3, NIPS 1990. Advances in Neural Information Processing Systems, vol. 3, Neural information processing systems foundation, pp. 356-362, 3rd Advances in Neural Information Processing Systems, NIPS 1990, Denver, United States, 26/11/90. https://doi.org/10.5555/118850.118898

Qualitative structure from motion. / Weinshall, Daphna.
Advances in Neural Information Processing Systems 3, NIPS 1990. ed. / Richard Lippmann; John E. Moody; David S. Touretzky. Neural information processing systems foundation, 1990. p. 356-362 (Advances in Neural Information Processing Systems; Vol. 3).

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

TY - GEN

T1 - Qualitative structure from motion

AU - Weinshall, Daphna

PY - 1990

Y1 - 1990

N2 - Exact structure from motion is an ill-posed computation and therefore very sensitive to noise. In this work I describe how a qualitative shape representation, based on the sign of the Gaussian curvature, can be computed directly from motion disparities, without the computation of an exact depth map or the directions of surface normals. I show that humans can judge the curvature sense of three points undergoing 3D motion from two, three and four views with success rate significantly above chance. A simple RBF net has been trained to perform the same task.

AB - Exact structure from motion is an ill-posed computation and therefore very sensitive to noise. In this work I describe how a qualitative shape representation, based on the sign of the Gaussian curvature, can be computed directly from motion disparities, without the computation of an exact depth map or the directions of surface normals. I show that humans can judge the curvature sense of three points undergoing 3D motion from two, three and four views with success rate significantly above chance. A simple RBF net has been trained to perform the same task.

UR - https://www.scopus.com/pages/publications/105020987401

U2 - 10.5555/118850.118898

DO - 10.5555/118850.118898

M3 - ???researchoutput.researchoutputtypes.contributiontobookanthology.conference???

AN - SCOPUS:105020987401

T3 - Advances in Neural Information Processing Systems

SP - 356

EP - 362

BT - Advances in Neural Information Processing Systems 3, NIPS 1990

A2 - Lippmann, Richard

A2 - Moody, John E.

A2 - Touretzky, David S.

PB - Neural information processing systems foundation

T2 - 3rd Advances in Neural Information Processing Systems, NIPS 1990

Y2 - 26 November 1990 through 29 November 1990

ER -

Qualitative structure from motion

Abstract

Publication series

Conference

Bibliographical note

Access to Document

Other files and links

Fingerprint

Cite this