3D Dense Captioning on ReferIt3D Nr3D (test)
45.53C Score (0.5 IoU)Vote2Cap-DETR
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Vote2Cap-DETRLoss=SCST2023.01 | 45.53 | 26.88 | 25.43 | 54.76 | |
| Vote2Cap-DETRLoss=MLE2023.01 | 43.84 | 26.68 | 25.41 | 54.43 | |
| D3NetTraining objective=CIDEr, Speaker-Listener Architecture=True, Additional ScanNet Data=True2021.12 | 38.42 | 22.22 | 24.74 | 54.37 | |
| D3NetLoss=SCST2023.01 | 38.42 | 22.22 | 24.74 | 54.37 | |
| 3DJCGLoss=MLE2023.01 | 38.06 | 22.82 | 23.77 | 52.99 | |
| D3NetTraining objective=CIDEr, Speaker-Listener Architecture=True2021.12 | 37.35 | 21.4 | 24.1 | 54.14 | |
| D3NetTraining objective=CIDEr2021.12 | 36.79 | 21.12 | 23.91 | 53.83 | |
| D3NetTraining objective=MLE2021.12 | 33.85 | 20.7 | 23.13 | 53.38 | |
| D3NetLoss=MLE2023.01 | 33.85 | 20.7 | 23.13 | 53.38 | |
| SpaCap3dLoss=MLE2023.01 | 33.71 | 19.92 | 22.61 | 50.5 | |
| X-Trans2CapLoss=SCST2023.01 | 33.62 | 19.29 | 22.27 | 50 | |
| Scan2CapLoss=MLE2023.01 | 27.47 | 17.24 | 21.8 | 49.06 | |
| Scan2Cap2021.12 | 22.38 | 13.87 | 20.44 | 47.96 |