3D Dense Captioning on ScanRefer (test)
86.28CIDErVote2Cap-DETR
Evaluation Results
| Method | Links | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Vote2Cap-DETRLdes=SCST, additional 2D input=true, IoU threshold=0.252023.01 | 86.28 | — | — | — | — | — | — | — | — | — | 42.64 | 28.27 | 59.07 | |
| Vote2Cap-DETRLdes=SCST, additional 2D input=false, IoU threshold=0.252023.01 | 84.15 | — | — | — | — | — | — | — | — | — | 42.51 | 28.47 | 59.26 | |
| Vote2Cap-DETRLdes=SCST, additional 2D input=false, IoU threshold=0.502023.01 | 73.77 | — | — | — | — | — | — | — | — | — | 38.21 | 26.64 | 54.71 | |
| Vote2Cap-DETRLdes=MLE, additional 2D input=true, IoU threshold=0.252023.01 | 72.79 | — | — | — | — | — | — | — | — | — | 39.17 | 28.06 | 59.23 | |
| Vote2Cap-DETRLdes=MLE, additional 2D input=false, IoU threshold=0.252023.01 | 71.45 | — | — | — | — | — | — | — | — | — | 39.34 | 28.25 | 59.33 | |
| Vote2Cap-DETRLdes=SCST, additional 2D input=true, IoU threshold=0.502023.01 | 70.63 | — | — | — | — | — | — | — | — | — | 35.69 | 25.51 | 52.28 | |
| 3DJCGLdes=MLE, additional 2D input=true, IoU threshold=0.252023.01 | 64.7 | — | — | — | — | — | — | — | — | — | 40.17 | 27.66 | 59.23 | |
| SpaCap3dLdes=MLE, additional 2D input=true, IoU threshold=0.252023.01 | 63.3 | — | — | — | — | — | — | — | — | — | 36.46 | 26.71 | 55.71 | |
| MORELdes=MLE, additional 2D input=true, IoU threshold=0.252023.01 | 62.91 | — | — | — | — | — | — | — | — | — | 36.25 | 26.75 | 56.33 | |
| D3NetLdes=SCST, additional 2D input=true, IoU threshold=0.502023.01 | 62.64 | — | — | — | — | — | — | — | — | — | 35.68 | 25.72 | 53.9 | |
| Scan2CapLdes=SCST, additional 2D input=true, IoU threshold=0.252023.01 | 61.83 | — | — | — | — | — | — | — | — | — | 35.65 | 26.61 | 54.7 | |
| Vote2Cap-DETRLdes=MLE, additional 2D input=false, IoU threshold=0.502023.01 | 61.81 | — | — | — | — | — | — | — | — | — | 34.46 | 26.22 | 54.4 | |
| 3DJCGLdes=MLE, additional 2D input=false, IoU threshold=0.252023.01 | 60.86 | — | — | — | — | — | — | — | — | — | 39.67 | 27.45 | 59.02 | |
| Vote2Cap-DETRLdes=MLE, additional 2D input=true, IoU threshold=0.502023.01 | 59.32 | — | — | — | — | — | — | — | — | — | 32.42 | 25.28 | 52.53 | |
| MORELdes=MLE, additional 2D input=false, IoU threshold=0.252023.01 | 58.89 | — | — | — | — | — | — | — | — | — | 35.41 | 26.36 | 55.41 | |
| X-Trans2CapLdes=SCST, additional 2D input=false, IoU threshold=0.252023.01 | 58.81 | — | — | — | — | — | — | — | — | — | 34.17 | 25.81 | 54.1 | |
| SpaCap3dLdes=MLE, additional 2D input=false, IoU threshold=0.252023.01 | 58.06 | — | — | — | — | — | — | — | — | — | 35.3 | 26.16 | 55.03 | |
| Scan2CapLdes=MLE, additional 2D input=true, IoU threshold=0.252023.01 | 56.82 | — | — | — | — | — | — | — | — | — | 34.18 | 26.29 | 55.27 | |
| Scan2CapLdes=MLE, additional 2D input=false, IoU threshold=0.252023.01 | 53.73 | — | — | — | — | — | — | — | — | — | 34.25 | 26.14 | 54.95 | |
| 3DJCGLdes=MLE, additional 2D input=true, IoU threshold=0.502023.01 | 49.48 | — | — | — | — | — | — | — | — | — | 31.03 | 24.22 | 50.8 | |
| 3DJCGLdes=MLE, additional 2D input=false, IoU threshold=0.502023.01 | 47.68 | — | — | — | — | — | — | — | — | — | 31.53 | 24.28 | 51.8 | |
| D3NetLdes=MLE, additional 2D input=true, IoU threshold=0.502023.01 | 46.07 | — | — | — | — | — | — | — | — | — | 30.29 | 24.35 | 51.67 | |
| SpaCap3dLdes=MLE, additional 2D input=true, IoU threshold=0.502023.01 | 44.02 | — | — | — | — | — | — | — | — | — | 25.26 | 22.33 | 45.36 | |
| Scan2CapLdes=SCST, additional 2D input=true, IoU threshold=0.502023.01 | 43.87 | — | — | — | — | — | — | — | — | — | 25.05 | 22.46 | 45.28 | |
| SpaCap3dLdes=MLE, additional 2D input=false, IoU threshold=0.502023.01 | 42.76 | — | — | — | — | — | — | — | — | — | 25.38 | 22.84 | 45.66 | |
| X-Trans2CapLdes=SCST, additional 2D input=false, IoU threshold=0.502023.01 | 41.52 | — | — | — | — | — | — | — | — | — | 23.83 | 21.9 | 44.97 | |
| MORELdes=MLE, additional 2D input=true, IoU threshold=0.502023.01 | 40.94 | — | — | — | — | — | — | — | — | — | 22.93 | 21.66 | 44.42 | |
| Scan2CapLdes=MLE, additional 2D input=true, IoU threshold=0.502023.01 | 39.08 | — | — | — | — | — | — | — | — | — | 23.32 | 21.97 | 44.78 | |
| MORELdes=MLE, additional 2D input=false, IoU threshold=0.502023.01 | 38.98 | — | — | — | — | — | — | — | — | — | 23.01 | 21.65 | 44.33 | |
| Scan2CapLdes=MLE, additional 2D input=false, IoU threshold=0.502023.01 | 35.2 | — | — | — | — | — | — | — | — | — | 22.36 | 21.44 | 43.57 | |
| 2D-3D Proj.Extra 2D=true, Proposals=Mask R-CNN2022.03 | — | 18.29 | 10.27 | 16.67 | 33.63 | 8.31 | 2.31 | 12.54 | 25.93 | 10.5 | — | — | — | |
| 3D-2D Proj.Extra 2D=true, Proposals=VoteNet2022.03 | — | 19.73 | 17.86 | 19.83 | 40.68 | 11.47 | 8.56 | 15.73 | 31.65 | 31.83 | — | — | — | |
| D3NetObjective=MLE2021.12 | — | — | — | — | — | 46.07 | 30.29 | 24.35 | 51.67 | — | — | — | — | |
| D3NetObjective=CIDEr2021.12 | — | — | — | — | — | 57.88 | 32.64 | 24.86 | 52.26 | — | — | — | — | |
| D3NetObjective=CIDEr, Localization=fixed2021.12 | — | — | — | — | — | 58.93 | 33.36 | 25.12 | 52.62 | — | — | — | — | |
| D3NetObjective=CIDEr, Localization=learned2021.12 | — | — | — | — | — | 61.3 | 34.5 | 25.25 | 52.8 | — | — | — | — | |
| D3NetObjective=CIDEr, Localization=learned, Auxiliary Loss=lobjcls2021.12 | — | — | — | — | — | 61.5 | 35.05 | 25.48 | 53.31 | — | — | — | — | |
| D3NetExtra Training Data=0.1x2021.12 | — | — | — | — | — | 61.91 | 35.03 | 25.38 | 53.25 | — | — | — | — | |
| D3NetExtra Training Data=0.5x2021.12 | — | — | — | — | — | 62.36 | 35.54 | 25.43 | 53.67 | — | — | — | — | |
| D3NetExtra Training Data=1x2021.12 | — | — | — | — | — | 62.64 | 35.68 | 25.72 | 53.9 | — | — | — | — | |
| Scan2capExtra 2D=false, Proposals=VoteNet2022.03 | — | 50.71 | 33.01 | 25.47 | 53.6 | 33.53 | 21.58 | 21.04 | 43.03 | 32.46 | — | — | — | |
| Scan2capExtra 2D=true, Proposals=VoteNet2022.03 | — | 56.82 | 34.18 | 26.29 | 55.27 | 39.08 | 23.32 | 21.97 | 44.78 | 32.21 | — | — | — | |
| Scan2Cap2021.12 | — | — | — | — | — | 39.08 | 23.32 | 21.97 | 44.78 | — | — | — | — | |
| TransCapExtra 2D=false, Proposals=VoteNet2022.03 | — | 55.36 | 32.46 | 25.64 | 53.19 | 40.08 | 22.86 | 21.72 | 44.04 | 33.34 | — | — | — | |
| TransCapExtra 2D=true, Proposals=VoteNet2022.03 | — | 60.04 | 35.04 | 26.27 | 54.46 | 43.12 | 24.25 | 22.15 | 44.72 | 34.34 | — | — | — | |
| X-Trans2CapExtra 2D=false, Proposals=VoteNet2022.03 | — | 58.81 | 34.17 | 25.81 | 54.1 | 41.52 | 23.83 | 21.9 | 44.97 | 34.68 | — | — | — | |
| X-Trans2CapExtra 2D=true, Proposals=VoteNet2022.03 | — | 61.83 | 35.65 | 26.61 | 54.7 | 43.87 | 25.05 | 22.46 | 45.28 | 35.31 | — | — | — | |
| X-Trans2Cap2021.12 | — | — | — | — | — | 43.87 | 25.05 | 22.46 | 44.97 | — | — | — | — |