List your company

3D human body pose and mesh estimation on SURREAL (test)

36.9MPJPEVirtual Marker
28.05687.753147.45207.147Aug 9, 2020Jan 15, 2021Jun 23, 2021Nov 29, 2021May 7, 2022Oct 13, 2022Mar 21, 2023

Evaluation results

MethodLinks
Virtual MarkerIntermediate Representation=Virtual marker2023.0336.9—28.944.7
I2L-MeshNet2020.0837.744.7——
BodyNet2020.0840.865.8——
PC-HMRbase architecture=HMR, 2D pose estimator=CPN, 2D-to-3D pose lifter=VideoPose3D2021.0351.7—37.959.8
PC-HMRIntermediate Representation=3D skeleton2023.0351.7—37.959.8
†DecoMRDtrain=true, Dtest=false2021.1152—4368.9
DecoMRIntermediate Representation=IUV image2023.0352—4368.9
DynaBOADtrain=false, Dtest=true2021.1155.2—3470.7
DynaBOAIntermediate Representation=-, training_on_test_split_with_2D_supervisions=true2023.0355.2—3470.7
Pose2MeshIntermediate Representation=3D skeleton2023.0356.6—39.668.8
Self-mocap2021.0364.4——74.5
†SPINDtrain=true, Dtest=false2021.1166.7—43.782.3
SPINIntermediate Representation=-2023.0366.7—43.782.3
HMR2021.0373.6—55.485.1
HMRIntermediate Representation=-2023.0373.6—55.485.1
*SPINDtrain=false, Dtest=true, Supervision=2D supervisions2021.1180.7—51.999.3
†GraphCMRDtrain=true, Dtest=false2021.1187.4—63.2103.2
GraphCMRIntermediate Representation=3D vertices2023.0387.4—63.2103.2
MRP-Net + ISOAdaptation type=Test-time Adaptation, Inference stage optimization=true2022.0396.4—65.1—
MRP-NetAdaptation type=General Adaptation, Inference stage optimization=false2022.0399.6—67.2—
Zhang et al.Adaptation type=General Adaptation, Inference stage optimization=false2022.03103.3—69.1—
SPINDtrain=false, Dtest=false2021.11105.6—68.6133.7
DANNAdaptation type=General Adaptation, Inference stage optimization=false2022.03113.6—77.2—
DANAdaptation type=General Adaptation, Inference stage optimization=false2022.03114.2—78.4—
DDCAdaptation type=General Adaptation, Inference stage optimization=false2022.03117.5—80.1—
GraphCMRDtrain=false, Dtest=false2021.11127.4—72150.8
DecoMRDtrain=false, Dtest=false2021.11138—74.9161
*GraphCMRDtrain=false, Dtest=true, Supervision=2D supervisions2021.11167.2—89.4189.2
*DecoMRDtrain=false, Dtest=true, Supervision=3D supervisions2021.11175.8—116.4227.4
Base modelDtrain=false, Dtest=false2021.11258—143.1278.6
Bodynet2021.03———73.6
BodyNetIntermediate Representation=Skel. + Seg.2023.03———65.8
SMPLify2020.08—75.3——