HD-EPIC
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| 6-DOF Object Trajectory Synthesis | HD-EPIC | ADE (m)0.279 | 11 | |
| Visual Question Answering | HD-EPIC | Recipe Accuracy65 | 10 | |
| Video Question Answering | HD-EPIC 1 (test) | Recipe96.7 | 9 | |
| Egocentric 4D Reasoning | HD-EPIC | Average Multiple-Choice Accuracy91.1 | 8 | |
| Audio-to-Text Retrieval | HD-EPIC | mAP32.2 | 8 | |
| Text-to-Audio Retrieval | HD-EPIC | mAP10.7 | 8 | |
| Parallel task execution prediction | HD-EPIC 2-Body Problem | Coverage1 | 8 | |
| Motion Generation | HD-EPIC curated P&R sequences | Prime Success66.31 | 8 | |
| Video Question Answering | HD-EPIC VQA | Recipe Accuracy88 | 7 | |
| Natural Language Query | HD-EPIC QnF | R@1 (IoU=0.3)15.3 | 7 | |
| Interaction Anticipation | HD-EPIC | Accuracy27.5 | 7 | |
| Egocentric Long Video Understanding | HD-EPIC++ | Accuracy30.28 | 7 | |
| Egocentric QA | HD-EPIC (test) | Accuracy46.2 | 6 | |
| Video Question Answering | HD-EPIC VQA 1.0 (test) | Following Activity Recognition84 | 5 | |
| Articulation Estimation | HD-EPIC 64 (test) | Match (%)71.43 | 5 | |
| Parallel task execution prediction | HD-EPIC 3-Body Problem | Coverage85.9 | 1 |
Showing 16 of 16 rows