Semantic Segmentation On Densepass
Métriques
mIoU
Résultats
Résultats de performance de divers modèles sur ce benchmark
Tableau comparatif
| Nom du modèle | mIoU |
|---|---|
| pyramid-vision-transformer-a-versatile | 31.20% |
| issafe-improving-semantic-segmentation-in | 32.04% |
| taking-a-closer-look-at-domain-shift-category | 31.46% |
| erfnet-efficient-residual-factorized-convnet | 16.65% |
| seamless-scene-segmentation | 34.14% |
| behind-every-domain-there-is-a-shift-adapting | 57.23% |
| differential-treatment-for-stuff-and-things-a | 44.58% |
| segformer-simple-and-efficient-design-for | 38.5% |
| rethinking-semantic-segmentation-from-a | 35.6% |
| bending-reality-distortion-aware-transformers | 55.25% |
| behind-every-domain-there-is-a-shift-adapting | 56.45% |
| dual-attention-network-for-scene-segmentation | 28.5% |
| fast-scnn-fast-semantic-segmentation-network | 24.6% |
| capturing-omni-range-context-for | 43.02% |
| segformer-simple-and-efficient-design-for | 42.4% |
| universal-semi-supervised-semantic | 30.87% |
| universal-semi-supervised-semantic | 26.98% |
| prototypical-cross-domain-self-supervised | 53.83% |
| disentangled-non-local-neural-networks | 32.1% |
| as-mlp-an-axial-shifted-mlp-architecture-for | 42.05% |
| daformer-improving-network-architectures-and | 54.67% |
| bending-reality-distortion-aware-transformers | 56.38% |
| encoder-decoder-with-atrous-separable | 32.5% |
| real-time-semantic-segmentation-with-fast | 26.9% |
| transfer-beyond-the-field-of-view-dense | 41.99% |
| in-defense-of-pre-trained-imagenet | 25.67% |
| confidence-regularized-self-training | 31.67% |
| pyramid-scene-parsing-network | 29.5% |
| transfer-beyond-the-field-of-view-dense | 48.52% |
| panoptic-feature-pyramid-networks | 28.8% |
| rethinking-semantic-segmentation-from-a | 35.7% |
| dpt-deformable-patch-based-transformer-for | 36.50% |
| ds-pass-detail-sensitive-panoramic-annular | 23.66% |
| understanding-the-robustness-in-vision | 42.54% |
| metaformer-is-actually-what-you-need-for | 43.18% |
| cyclemlp-a-mlp-like-architecture-for-dense | 40.16% |