Motion integration using competitive priors

Shuang WU, Hongjing LU, Lap Fai, Alan LEE, Alan YUILLE

Research output: Book Chapters | Papers in Conference ProceedingsBook Chapter

1 Scopus Citations

Abstract

Psychophysical experiments show that humans are better at perceiving rotation and expansion than translation [5][9]. These findings are inconsistent with standard models of motion integration which predict best performance for translation. To explain this discrepancy, our theory formulates motion perception at two levels of inference: we first perform model selection between the competing models (e.g. translation, rotation, and expansion) and then estimate the velocity using the selected model. We define novel prior models for smooth rotation and expansion using techniques similar to those in the slow-and-smooth model [23] (e.g. Green functions of differential operators). The theory gives good agreement with the trends observed in four human experiments.
Original languageEnglish
Title of host publicationStatistical and geometrical approaches to visual motion analysis : International Dagstuhl Seminar, Dagstuhl Castle, Germany, July 13-18, 2008
PublisherSpringer
Pages235-258
Number of pages24
ISBN (Print)9783642030604
DOIs
Publication statusPublished - 1 Jan 2009
Externally publishedYes

Fingerprint

Green's function
Mathematical operators
Experiments

Cite this

WU, S., LU, H., LEE, L. F. A., & YUILLE, A. (2009). Motion integration using competitive priors. In Statistical and geometrical approaches to visual motion analysis : International Dagstuhl Seminar, Dagstuhl Castle, Germany, July 13-18, 2008 (pp. 235-258). Springer. https://doi.org/10.1007/978-3-642-03061-1_12
WU, Shuang ; LU, Hongjing ; LEE, Lap Fai, Alan ; YUILLE, Alan. / Motion integration using competitive priors. Statistical and geometrical approaches to visual motion analysis : International Dagstuhl Seminar, Dagstuhl Castle, Germany, July 13-18, 2008. Springer, 2009. pp. 235-258
@inbook{fdad87ae695b48489851b600f49c32ae,
title = "Motion integration using competitive priors",
abstract = "Psychophysical experiments show that humans are better at perceiving rotation and expansion than translation [5][9]. These findings are inconsistent with standard models of motion integration which predict best performance for translation. To explain this discrepancy, our theory formulates motion perception at two levels of inference: we first perform model selection between the competing models (e.g. translation, rotation, and expansion) and then estimate the velocity using the selected model. We define novel prior models for smooth rotation and expansion using techniques similar to those in the slow-and-smooth model [23] (e.g. Green functions of differential operators). The theory gives good agreement with the trends observed in four human experiments.",
author = "Shuang WU and Hongjing LU and LEE, {Lap Fai, Alan} and Alan YUILLE",
year = "2009",
month = "1",
day = "1",
doi = "10.1007/978-3-642-03061-1_12",
language = "English",
isbn = "9783642030604",
pages = "235--258",
booktitle = "Statistical and geometrical approaches to visual motion analysis : International Dagstuhl Seminar, Dagstuhl Castle, Germany, July 13-18, 2008",
publisher = "Springer",

}

WU, S, LU, H, LEE, LFA & YUILLE, A 2009, Motion integration using competitive priors. in Statistical and geometrical approaches to visual motion analysis : International Dagstuhl Seminar, Dagstuhl Castle, Germany, July 13-18, 2008. Springer, pp. 235-258. https://doi.org/10.1007/978-3-642-03061-1_12

Motion integration using competitive priors. / WU, Shuang; LU, Hongjing; LEE, Lap Fai, Alan; YUILLE, Alan.

Statistical and geometrical approaches to visual motion analysis : International Dagstuhl Seminar, Dagstuhl Castle, Germany, July 13-18, 2008. Springer, 2009. p. 235-258.

Research output: Book Chapters | Papers in Conference ProceedingsBook Chapter

TY - CHAP

T1 - Motion integration using competitive priors

AU - WU, Shuang

AU - LU, Hongjing

AU - LEE, Lap Fai, Alan

AU - YUILLE, Alan

PY - 2009/1/1

Y1 - 2009/1/1

N2 - Psychophysical experiments show that humans are better at perceiving rotation and expansion than translation [5][9]. These findings are inconsistent with standard models of motion integration which predict best performance for translation. To explain this discrepancy, our theory formulates motion perception at two levels of inference: we first perform model selection between the competing models (e.g. translation, rotation, and expansion) and then estimate the velocity using the selected model. We define novel prior models for smooth rotation and expansion using techniques similar to those in the slow-and-smooth model [23] (e.g. Green functions of differential operators). The theory gives good agreement with the trends observed in four human experiments.

AB - Psychophysical experiments show that humans are better at perceiving rotation and expansion than translation [5][9]. These findings are inconsistent with standard models of motion integration which predict best performance for translation. To explain this discrepancy, our theory formulates motion perception at two levels of inference: we first perform model selection between the competing models (e.g. translation, rotation, and expansion) and then estimate the velocity using the selected model. We define novel prior models for smooth rotation and expansion using techniques similar to those in the slow-and-smooth model [23] (e.g. Green functions of differential operators). The theory gives good agreement with the trends observed in four human experiments.

UR - http://commons.ln.edu.hk/sw_master/4489

U2 - 10.1007/978-3-642-03061-1_12

DO - 10.1007/978-3-642-03061-1_12

M3 - Book Chapter

SN - 9783642030604

SP - 235

EP - 258

BT - Statistical and geometrical approaches to visual motion analysis : International Dagstuhl Seminar, Dagstuhl Castle, Germany, July 13-18, 2008

PB - Springer

ER -

WU S, LU H, LEE LFA, YUILLE A. Motion integration using competitive priors. In Statistical and geometrical approaches to visual motion analysis : International Dagstuhl Seminar, Dagstuhl Castle, Germany, July 13-18, 2008. Springer. 2009. p. 235-258 https://doi.org/10.1007/978-3-642-03061-1_12