Spatial Visual Question Answering on IaOR-VQA Qualitative
87.3AccuracyGPT-4o
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-4o3D information enhancement=w ours2024.03 | 87.3 | |
| GPT-4V3D information enhancement=w ours2024.03 | 86.3 | |
| LLaVA-1.53D information enhancement=w ours2024.03 | 83 | |
| InstructBLIP3D information enhancement=w ours2024.03 | 79.3 | |
| SpatialVLMimplementation=HF version2024.03 | 76.7 | |
| GPT-4V3D information enhancement=w/o ours2024.03 | 70.7 | |
| LLaVA-1.53D information enhancement=w/o ours2024.03 | 70 | |
| GPT-4o3D information enhancement=w/o ours2024.03 | 69 | |
| InstructBLIP3D information enhancement=w/o ours2024.03 | 62.3 |