Datasets:
tensorkelechi
/
arxiv-cstext

Dataset card Data Studio Files Community
1
text
stringlengths
0
89.3k
48 Haitian Zeng Xin Yu Jiaxu Miao and Yi Yang Mhrnet
Multiplehypothesis reconstruction of nonrigid shapes from
2d views In Eur Conf Comput Vis pages 117 2022 1 2
49 Yingying Zhu Dong Huang Fernando De La Torre and Si
mon Lucey Complex nonrigid motion 3d reconstruction
by union of subspaces In IEEE Conf Comput Vis Pattern
Recog pages 15421549 2014 1 2
50 Yue Zhu Nermin Samet and David Picard H3wb Human3
6m 3d wholebody dataset and benchmark In Int Conf Com
put Vis pages 2016620177 2023 7Nonrigid StructurefromMotion Temporallysmooth Procrustean Alignment
and Spatiallyvariant Deformation Modeling
Supplementary Material
Abstract
In this supplementary material we provide additional
materials in the following aspects 1 solution of the
Temporallysmooth Procrustean Alignment TPA model
and verification of its validity 2 detailed solution of the
ADMM model 3 additional explanations for spatially
variant deformation modeling and 4 additional details of
the experimental setup supplementation of missing data ex
periments and more visualization results
6 Temporallysmooth Procrustean Alignment
In this section we first provide details in solving our pro
posed Temporallysmooth Procrustean Alignment TPA
module and then report additional experimental results
61 Algorithm for TPA
Denote the 3D shape sequence to be aligned as SiF
i1 we
set the optimization target of TPA as
LQ
tpa1
2F1X
i1QiSiQi1Si12
F 14
where we initialize QiI3 andI3R33denotes the
identity matrix We solve this optimization problem in a
framebyframe manner When 1 i F we transfer the
optimization target as
LQi
tpa1
2iX
ki1PX
j1QkskjQk1sk1j2
2
1
2PX
j1Qi1si1jQisij2
2
QisijQi1si1j2
2 15
where sijis the jth column of 3D shape Si Taking the
Lie algebra ϕicorresponding to the rotation matrix Qias
the optimization variable the gradient of model 14 can be
calculated as
giLQi
tpa
ϕi1
2PP
j1
r1
ij
2
2
r0
ij
2
2
ϕi16PX
j1
r1
ij
ϕiTr1
ij r0
ij
ϕiTr0
ij
PX
j1
QisijT
r1
ijr0
ij
where rm
ijQimsimjQim1sim1jis the resid
ual vector and Qisijis an approximation to the deriva
tive of the residual rm
ijwith respect to the Lie algebra ϕi
ais the skewsymmetric matrix of vector aie
a1