Image To Image Translation On Cityscapes
Métriques
FID
Per-pixel Accuracy
mIoU
Résultats
Résultats de performance de divers modèles sur ce benchmark
Tableau comparatif
| Nom du modèle | FID | Per-pixel Accuracy | mIoU |
|---|---|---|---|
| semi-parametric-image-synthesis | 49.7 | 75.5% | 47.2 |
| wavelet-based-unsupervised-label-to-image-1 | 50.14 | - | 42.32 |
| semantic-image-synthesis-with-spatially | 71.8 | 81.9% | 62.3 |
| diverse-semantic-image-synthesis-via | - | - | - |
| usis-unsupervised-semantic-image-synthesis | 53.67 | - | 44.78 |
| image-to-image-translation-with-conditional | - | 71.0 | - |
| improving-augmentation-and-evaluation-schemes | 52.1 | - | 63.1 |
| sesame-semantic-editing-of-scenes-by-adding | 54.2 | 82.5% | 66 |
| unlocking-pre-trained-image-backbones-for | 38.2 | - | 76.3 |
| high-resolution-image-synthesis-and-semantic | 95 | 81.4% | 58.3 |
| you-only-need-adversarial-supervision-for-1 | 47.7 | - | 69.3 |
| adversarially-learned-inference | - | 19% | - |
| coupled-generative-adversarial-networks | - | 40% | - |
| photographic-image-synthesis-with-cascaded | 104.7 | 77.1% | 52.4 |
| semantic-bottleneck-scene-generation | 60.39 | - | - |
| dual-pyramid-generative-adversarial-networks | 44.1 | - | 73.6 |
| learning-from-simulated-and-unsupervised | - | 20% | - |
| learning-to-predict-layout-to-image | 54.3 | 82.3% | 65.5 |
| improving-augmentation-and-evaluation-schemes | 72.7 | - | 58 |
| focal-frequency-loss-for-generative-models | 59.5 | 82.5% | 64.2 |
| unpaired-image-to-image-translation-using | - | 52% | - |