List your company

Novel View Synthesis on ZJU-MoCap (test)

0.977SSIMGART
0.692040.766020.840.91398Dec 21, 2021Aug 23, 2022Apr 25, 2023Dec 27, 2023Aug 28, 2024Apr 30, 2025Jan 1, 2026

Evaluation results

MethodLinks
GARTTraining time=~2.5min2023.110.977—32.2229.21————
GARTTraining time=~30s2023.110.976—31.7634.01————
ASTraining Time=~10 h2023.020.975—30.3837.23————
ASTraining time=~10h2023.110.975—30.3837.23————
ASTrain=10h, FPS=0.402023.120.975—30.3837.23————
RigGSTemporal sampling density=Full monocular video sequences2026.010.975—33.54————3.27
Instant-NVRTraining Time=~5 min2023.020.971—31.0138.45————
Instant-NVRTraining time=~5min2023.110.971—31.0138.45————
InstantNVRTrain=5m, FPS=2.202023.120.971—31.0138.45————
HumanNeRFTraining Time=~10 h2023.020.969—30.6633.38————
HumanNeRFTraining time=~10h2023.110.969—30.6633.38————
HumanNeRFTrain=10h, FPS=0.302023.120.969—30.6633.38————
Watch It MoveMerge strategy=without merge2021.120.9660.053——————
Watch It MoveMerge strategy=with merge2021.120.9650.056——————
ANTraining Time=~10 h2023.020.965—29.7746.89————
ANTraining time=~10h2023.110.965—29.7746.89————
ANTrain=10h, FPS=1.112023.120.965—29.7746.89————
GauHumanTrain=1m, FPS=1892023.120.965—31.3430.51————
NBTraining Time=~10 h2023.020.964—29.0342.47————
NBTraining time=~10h2023.110.964—29.0342.47————
NBTrain=10h, FPS=1.482023.120.964—29.0342.47————
dynamic MLP mapsImage resolution=512 x 5122023.040.9630.06830.17—24245——
C-NeRFImage resolution=512 x 5122023.040.9590.07729.79—1,3131,019——
DyNeRFImage resolution=512 x 5122023.040.9590.08729.88—3,45212——
Kundu* et al.adaptation=overfits SMPL parameters with MLP2021.120.9570.059——————
Schmidtke* et al.adaptation=deforms 3D template with MLP2021.120.9570.055——————
DVATrain=1.5h, FPS=16.52023.120.956—29.4537.74————
NHPTraining Time=~1 h fine-tuning, Pretraining Time=10 h2023.020.955—28.2564.77————
NHPTraining time=~1h tuning2023.110.955—28.2564.77————
NHPTrain=1h+, FPS=0.152023.120.955—28.2564.77————
SV-GSTemporal sampling density=5x fewer frames2026.010.944—28.83————5.89
InstantAvatarTrain=3m, FPS=4.152023.120.938—29.7368.41————
NVImage resolution=512 x 5122023.040.9340.13128.12—49658——
SV-GSTemporal sampling density=10x fewer frames2026.010.934—28.13————6.53
D-NeRFImage resolution=512 x 5122023.040.9220.13927.08—1,5344——
AP-NeRFTemporal sampling density=Full monocular video sequences2026.010.919—25.62————9.34
PixelNeRFTraining Time=~1 h fine-tuning, Pretraining Time=10 h2023.020.892—24.71121.86————
PixelNeRFTraining time=~1h tuning2023.110.892—24.71121.86————
PixelNeRFTrain=1h+, FPS=1.202023.120.892—24.71121.86————
SurMoCropped to 2D bounding boxes=true2024.040.8330.07524.815———67.725—
HumanNeRFCropped to 2D bounding boxes=true2024.040.7920.10623.624———88.382—
Instant-NVRCropped to 2D bounding boxes=true2024.040.7380.20221.748———144.99—
Neural BodyCropped to 2D bounding boxes=true2024.040.7030.16421.438———125.1—