A Hierarchical Static-Dynamic Encoder-Decoder Structure for 3D Human Motion Prediction with Residual CNNs

<div>Jointwise performance of different methods. (a) Jointwise MSE and (b) Jointwise MAE of different methods on G3D. (c) Jointwise MSE and (d) Jointwise MAE of different methods on FNTU.</div>

Mathematical Problems in Engineering

fig7

Figure 7

Figure 7: A Hierarchical Static-Dynamic Encoder-Decoder Structure for 3D Human Motion Prediction with Residual CNNs 