|Table of Contents|

Three-dimensional Modeling and Motion Estimation of Elastic Articulated Objects from Image Contours


Research Field:
Publishing date:


Three-dimensional Modeling and Motion Estimation of Elastic Articulated Objects from Image Contours
PAN Hai-lang1SHI Lei2ZHU Min3DAI Yue-wei1
1.School of Automation,NUST,Nanjing 210094,China;2.Shanghai Berkeley Biology and Technology Limited Company,Shanghai 200433,China;3.College of Information Science and Technology,Nanjing University of Aeronautics and Astronautics,Nanjing 210016,China
elastic articulated objects human body modeling revolving conic surfaces motion estimation
This paper presents a new method of elastic articulated objects(human bodies) modeling based on a new revolving conic surface.The model includes a three-dimensional object skeleton and deformable surfaces equations which can represent the deformation of human body surfaces.The deformation of human body surfaces can be represented by adjusting two deformation parameters for each limb.The three-dimensional deformation and motion parameters are determined by corresponding two-dimensional image contours from a sequences of stereo images.The algorithm presented here includes motion parameter estimation of the skeleton and parameter determination of the deformable surfaces.


[ 1] M oeslund T B, H ilton A, Kr?ger V. A survey o f advances in v ision-based hum an mo tion capture and ana-l ysis[ J]. Com puter V ision and Im age Understand ing, 2006, 104( 2): 90- 126.
[ 2] NahasM, Hu itr icH, Sa intourensM. Anim a tion of a B-sp line figure[ J]. V isual Compu ter, 1988, 3 ( 5): 272- 276.
[ 3] Pe tland A, H orow itz B. Recovery o f nonrig id m otion and struc ture[ J]. IEEE T ransactions on Pa ttern Ana-l ys is and M ach ine Inte lligence, 1991, 13 ( 7) : 730 - 742.
[ 4] M in K H, Baek SM, Lee G A, et a.l Ana tom ica llybased m ode ling and anim a tion o f hum an upper lim bs [ A] . Proceedings of In ternational Con ference on H um an Mode ling and An im ation [ C ] . Pohang: IEEE Computer Soc ie ty Press, 2000. 1- 14.
[ 5] P lankers R, Fua P. A rticulated soft ob jects form ultiview shape and m o tion capture[ J] . IEEE Transactions on Patte rn Ana lysis and M ach ine Intellig ence, 2003, 25 ( 9) : 1182- 1187.
[ 6] Sm inch isescu C. Estmi ation a lgorithms fo r ambiguous v isua lmode ls- three dmi ensional hum an mode ling andm otion reconstruction in monocular v ideo sequences[ D]. Paris: The French National Institute fo r Research in Computer Science and Contro,l 2002. 35- 36.
[ 7] D. Apuzzo N, P l&ankers R, Gruen A, et a.l M odelling hum an bodies from v ideo sequences[ A ]. Proceed ing s o f E lectron ic Im ag ing [ C ]. San Jo se: SPIE Press, 1999. 1- 12.
[ 8] Ja in R C, Kastur i R, Schunk B. M ach ine V ision [M ]. New York: M cGraw-H ill Inc, 1995.
[ 9] Zhang X Y, L iu Y C, H uang T S. A rticulated jo int estim ation from mo tion us ing two m onocular im ag es [ J]. Pattern Recognition Letters, 2004, 25 ( 10): 1097- 1106.
[ 10] M a S D. Con ics-based stereo, mo tion estima tion, and pose de term ina tion [ J] . Intern J Com puter V is ion, 1993, 10 ( 1) : 7- 25.
[ 11] Zhang Z Y. Param ete r estim ation techniques: a tuto r-i a lw ith app lication to con ic fitting[ J]. Im age and V-i sion Com puting Journa,l 1997, 15( 1): 59- 76.
[ 12] Zhang X Y, L iu Y C, Huang T S. M otion analysis o f articu la ted ob jects from m onocu lar im ages[ J]. IEEE Transac tions on Pattern Ana ly sis andM ach ine In tell-i gence, 2006, 28( 4): 625- 663.
[ 13] Zhang Z Y. A flex ib le new techn ique fo r cam era ca l-i bra tion [ J]. IEEE Transactions on Pa ttern Analysis and M ach ine Inte llig ence, 2000, 22 ( 11 ): 1330 - 1334.


Last Update: 2012-11-19