Visual Question Answering on PartNet-Reasoning Table 1.0 (test)
80.3Accuracy (Part Existence)3D-CG
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| 3D-CGArchitecture=Neural-Symbolic2022.07 | 80.3 | 75.1 | 90.9 | |
| BAE+LArchitecture=Neural-Symbolic2022.07 | 71 | 73.6 | 45.7 | |
| CVX+LArchitecture=Neural-Symbolic2022.07 | 68.5 | 69.7 | 30.7 | |
| MACArchitecture=Neural Network2022.07 | 66.4 | 51.2 | 55.3 | |
| NDF+LSTMArchitecture=Neural Network2022.07 | 61.3 | 54.4 | 50.1 | |
| PointNet+LSTMArchitecture=Neural Network2022.07 | 55.1 | 43.6 | 35.5 |