Table 5 Quantitative results of different models on Radar echo dataset.
From: 3D long time spatiotemporal convolution for complex transfer sequence prediction
Methods | CSI\(\uparrow\) | POD\(\uparrow\) | FAR\(\downarrow\) | MSE\(\downarrow\) | SSIM | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|
t>15 | t>25 | t>35 | t>15 | t>25 | t>35 | t>15 | t>25 | t>35 | |||
ConvLSTM | 0.57 | 0.50 | 0.22 | 0.70 | 0.55 | 0.32 | 0.22 | 0.19 | 0.18 | 19.21 | 0.79 |
VPN | 0.58 | 0.51 | 0.24 | 0.72 | 0.56 | 0.33 | 0.21 | 0.18 | 0.17 | 17.82 | 0.81 |
PredRNN | 0.61 | 0.51 | 0.25 | 0.73 | 0.57 | 0.34 | 0.19 | 0.18 | 0.15 | 17.01 | 0.81 |
MIM | 0.62 | 0.53 | 0.26 | 0.73 | 0.58 | 0.36 | 0.18 | 0.16 | 0.15 | 16.32 | 0.82 |
E3D-LSTM | 0.62 | 0.53 | 0.28 | 0.75 | 0.58 | 0.36 | 0.18 | 0.17 | 0.14 | 15.93 | 0.83 |
STAM | 0.63 | 0.54 | 0.30 | 0.76 | 0.60 | 0.38 | 0.16 | 0.15 | 0.13 | 15.55 | 0.82 |
Ours | 0.64 | 0.55 | 0.32 | 0.77 | 0.63 | 0.40 | 0.18 | 0.17 | 0.15 | 15.52 | 0.84 |