Our Dataset
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| AI-generated video detection | Our Dataset HunyuanVideo | Top-1 Accuracy86.17 | 16 | |
| AI-generated video detection | Our Dataset CogVideoX | Top-1 Accuracy84.32 | 16 | |
| Homography Estimation | Our Dataset | RE12.75 | 12 | |
| Image Editing | our dataset film-grey style | PSNR24.14 | 11 | |
| Novel View Synthesis | Our dataset real | PSNR23.46 | 8 | |
| Human Animation | Our Dataset | Pose Confidence91.1 | 7 | |
| Scene Synthesis | Our Dataset | FID27.3 | 5 | |
| Inertial Odometry | Our Dataset (test) | FDE5.9 | 5 | |
| Human Image Animation | Our dataset (test) | Subject Consistency92.54 | 5 | |
| Video Virtual Try-On | Our Dataset Internet Videos v1 (test) | FID71 | 5 | |
| Controllable stroke generation | Our dataset oil painting strokes 1.0 (test) | LPIPS0.232 | 4 | |
| Material Selection | Our Dataset (test) | mIoU69 | 4 | |
| Video Virtual Try-On | Our Dataset (test) | Video Smoothness0.02 | 4 | |
| Panoptic segmentation | Our dataset | PQ47.1 | 4 | |
| Hand-object manipulation video generation | Our Dataset | FID37.7 | 3 | |
| View Synthesis | Our dataset Hard Set 1.0 m to 2.0 m | Human Preference Rate0.73 | 3 | |
| View Synthesis | Our dataset Easy Set (0.2 m to 0.3 m) | Human Preference Rate66 | 3 | |
| Malicious Smart Contract Detection | Our Dataset | TPR99 | 2 | |
| 3D Multi-person Pose Estimation | Our Dataset (test) | Precision88.5 | 2 | |
| Panoptic Segmentation | Our Dataset (test) | PQ47.1 | 2 | |
| Panoptic Segmentation | Our Dataset (inference) | Inference Latency (ms)264 | 2 |
Showing 21 of 21 rows