Panoptic Segmentation on PASTIS
81.5SQU-TAE + PaPs
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| U-TAE + PaPsS=16, temporal_input=multi-temporal, shape_prediction=residual CNN2021.07 | 81.5 | 53.2 | 43.8 | |
| U-TAE + PaPsS=8, temporal_input=multi-temporal2021.07 | 80.9 | 52.3 | 42.7 | |
| U-TAE + PaPsS=24, temporal_input=multi-temporal2021.07 | 80.7 | 50.6 | 41.3 | |
| U-ConvLSTM + PaPsS=16, temporal_input=multi-temporal2021.07 | 80.2 | 43.9 | 35.6 | |
| U-TAE + PaPsS=16, shape_prediction=Multiplicative Saliency2021.07 | 74.6 | 49.9 | 37.5 | |
| U-TAE + PaPsS=16, temporal_input=Single-image (August)2021.07 | 72.3 | 18.7 | 14.1 |