Taskonomy
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Depth Estimation | Taskonomy (test) | Depth Estimation Error0.022 | 21 | |
| Keypoint Detection | Taskonomy tiny (test) | MAE0.191 | 16 | |
| Semantic Segmentation | Taskonomy tiny (test) | mIoU73.7 | 16 | |
| Semantic Segmentation | Taskonomy (test) | mIoU73.7 | 16 | |
| Edge Detection | Taskonomy (test) | Edge Det.21.7 | 10 | |
| Keypoint Detection | Taskonomy (test) | Keypoint Detection20.2 | 10 | |
| Surface Normal Prediction | Taskonomy (test) | Surface Normal Accuracy87.3 | 10 | |
| Edge Detection | Taskonomy | MAE0.046 | 9 | |
| Surface Normal Estimation | Taskonomy | CS0.931 | 9 | |
| Depth Estimation | Taskonomy | MAE0.013 | 9 | |
| Segmentation | Taskonomy | CE0.462 | 9 | |
| Multi-task learning | Tiny-Taskonomy (test) | Delta t1 (Semantic Segmentation)-28.2 | 9 | |
| Edge Detection | Taskonomy Tiny (test) | Edge Score21.7 | 8 | |
| Depth Prediction | Taskonomy Tiny (test) | Depth Error0.022 | 8 | |
| Image Inpainting | Taskonomy tiny (test) | FID59.3 | 8 | |
| Scene Recognition | Taskonomy tiny (test) | Accuracy71 | 8 | |
| Multimodal Image-to-Image Translation (RGB+Edge to Depth) | Taskonomy | MAE0.16 | 8 | |
| Multimodal Image-to-Image Translation (RGB+Normal to Shade) | Taskonomy | MAE0.79 | 8 | |
| Multimodal Image-to-Image Translation (RGB+Shade to Normal) | Taskonomy | MAE0.58 | 8 | |
| Multimodal Image-to-Image Translation (Depth+Normal to RGB) | Taskonomy | FID70.13 | 8 | |
| Multimodal Image-to-Image Translation (Shade+Texture to RGB) | Taskonomy | FID43.92 | 8 | |
| Surface Normal Prediction | Tiny-Taskonomy | SN (Accuracy)0.707 | 8 | |
| Multi-task Learning | Taskonomy tiny (test) | Object Classification89.75 | 7 | |
| Surface Normal Estimation | Taskonomy AE (test) | L1 Error4.94 | 7 | |
| Surface Normal Estimation | Taskonomy 3DCC (test) | L1 Error5.35 | 7 |
Showing 25 of 47 rows