Physically Plausible Animation of Human Upper Body from a Single Image Visualization

Frame by frame evaluation result of Physically Plausible Animation of Human Upper Body from a Single Image (Ours), Controllable Video Generation with Sparse Trajectories(Hao et al.), 3D Simulation, and Motion Reconstruction Code and Data for Skills from Videos (SFV). For each sample, the top-left frame is the ground truth video. All methods aim to generate a video in which the person’s wrists move along the ground-truth wrist trajectories. The top-second-to-last videos are the synthesized videos generated by our method, Hao et al., 3D simulation, SfV respectively. The bottom-left video overlays the predicted poses generates by our method over our result. The other bottom videos show the per-pixel PSNR values of the top corresponding method as heatmaps. For PSNR, higher is better. For LPIPS, lower is better. Our results generally look more plausible and sharper than the baselines’, as reflected by LPIPS. However, the misalignment of the head and arms in our results lead to a worse PSNR score than the blurry results of Hao et al.'s method. Note that SfV requires the whole ground truth video as the input for training, while other method only takes the first frame and the 2D wrist trajectories of the ground truth video as input.

Total Result

MethodOursHao et al.3D SimulationSfV
PSNR 22.633 24.308 21.230 22.227
LPIPS 0.121 0.134 0.160 0.130

sample 1


MethodOursHao et al.3D SimulationSfV
PSNR 24.11725.61921.46322.517
LPIPS 0.0790.0830.1330.099

sample 2


MethodOursHao et al.3D SimulationSfV
PSNR 23.11623.67320.82820.442
LPIPS 0.1040.1410.1840.166

sample 3


MethodOursHao et al.3D SimulationSfV
PSNR 21.48622.56619.65522.813
LPIPS 0.1310.1620.2130.113

sample 4


MethodOursHao et al.3D SimulationSfV
PSNR 21.98622.40322.15421.828
LPIPS 0.1920.1720.1490.197

sample 5


MethodOursHao et al.3D SimulationSfV
PSNR 22.71023.69121.62222.354
LPIPS 0.1060.1340.1460.126

sample 6


MethodOursHao et al.3D SimulationSfV
PSNR 22.97622.74922.08921.699
LPIPS 0.1210.1800.1450.166

sample 7


MethodOursHao et al.3D SimulationSfV
PSNR 23.96124.11921.61222.762
LPIPS 0.1000.1370.1360.096

sample 8


MethodOursHao et al.3D SimulationSfV
PSNR 23.35127.50321.76722.388
LPIPS 0.1010.0900.1330.120

sample 9


MethodOursHao et al.3D SimulationSfV
PSNR 23.38827.51720.78823.343
LPIPS 0.0840.0530.1600.086

sample 10


MethodOursHao et al.3D SimulationSfV
PSNR 22.84823.84221.25222.118
LPIPS 0.1040.1460.1540.130