Semantic Segmentation On Nyu Depth V2
Métriques
Mean IoU
Résultats
Résultats de performance de divers modèles sur ce benchmark
Tableau comparatif
| Nom du modèle | Mean IoU |
|---|---|
| dynamic-multimodal-fusion | 51.0% |
| geminifusion-efficient-pixel-wise-multimodal | 60.9 |
| cmx-cross-modal-fusion-for-rgb-x-semantic | 54.4% |
| locality-sensitive-deconvolution-networks | 45.9% |
| efficient-rgb-d-semantic-segmentation-for | 48.17 |
| 3d-graph-neural-networks-for-rgbd-semantic | 43.1% |
| asymformer-asymmetrical-cross-modal | 55.3% |
| learning-deep-multimodal-feature | 51.2% |
| swinmtl-a-shared-architecture-for | 58.14% |
| omnivec2-a-novel-transformer-based-network | 63.6 |
| multi-modal-attention-based-fusion-model-for | 44.8% |
| geminifusion-efficient-pixel-wise-multimodal | 56.8 |
| cross-stitch-networks-for-multi-task-learning | 19.3% |
| rgb-based-semantic-segmentation-using-self | 33.49% |
| delivering-arbitrary-modal-semantic | 56.9% |
| cascaded-feature-network-for-semantic | 47.7% |
| nddr-cnn-layer-wise-feature-fusing-in-multi | 43.3% |
| acnet-attention-based-network-to-exploit | 48.3% |
| malleable-2-5d-convolution-learning-receptive | 50.9% |
| spatial-information-guided-adaptive-context | 49.4% |
| adashare-learning-what-to-share-for-efficient | 29.6% |
| pixel-difference-convolutional-network-for | 53.5% |
| optimizing-rgb-d-semantic-segmentation | 51.9% |
| masked-supervised-learning-for-semantic | 39.31% |
| multimodal-token-fusion-for-vision | 53.3% |
| haarnet-large-scale-linear-morphological | 50.7% |
| improving-multi-modal-learning-with-uni-modal | 49.14% |
| exploring-relational-context-for-multi-task | 46.33% |
| understanding-dark-scenes-by-contrasting | 52.5% |
| composite-learning-for-robust-and-effective | 33.48% |
| refinenet-multi-path-refinement-networks-for | 46.5% |
| inverseform-a-loss-function-for-structured | 53.1% |
| ci-net-contextual-information-for-joint | 42.6% |
| prompt-guided-transformer-for-multi-task | 41.61 |
| comptr-towards-diverse-bi-source-dense | 55.5% |
| efficient-yet-deep-convolutional-neural | 32.3% |
| multi-task-meta-learning-learn-how-to-adapt | 41.51% |
| variational-context-deformable-convnets-for | 50.7% |
| omnivore-a-single-model-for-many-visual | 55.1% |
| joint-task-recursive-learning-for-semantic | 46.8% |
| dcanet-differential-convolution-attention | 53.3% |
| inverted-pyramid-multi-task-transformer-for | 53.56% |
| toward-edge-efficient-dense-predictions-with | 22.1% |
| multimodal-knowledge-expansion | 48.88% |
| efficient-multi-task-scene-analysis-with-rgb | 51.26% |
| malleable-2-5d-convolution-learning-receptive | 49.7% |
| mmanet-margin-aware-distillation-and-modality | 49.62% |
| learning-fully-dense-neural-networks-for | 47.4% |
| multimodal-token-fusion-for-vision | 54.2% |
| mti-net-multi-scale-task-interaction-networks | 49.0 |
| pattern-structure-diffusion-for-multi-task | 51.0% |
| multi-layer-feature-aggregation-for-deep | 50.7% |
| light-weight-refinenet-for-real-time-semantic | 44.4% |
| temporally-distributed-networks-for-fast | 43.5 |
| understanding-dark-scenes-by-contrasting | 55.8% |
| multimae-multi-modal-multi-task-masked | 56.0% |
| hs3-learning-with-proper-task-complexity-in | 53.5% |
| bi-directional-cross-modality-feature | 52.4% |
| scene-parsing-via-integrated-classification | 50.70 |
| polymax-general-dense-prediction-with-mask | 58.08% |
| mmformer-multimodal-medical-transformer-for | 48.45% |
| rednet-residual-encoder-decoder-network-for | 47.2% |
| variational-context-deformable-convnets-for | 51.9% |
| dense-decoder-shortcut-connections-for-single | 48.1% |
| efficient-multimodal-semantic-segmentation | 59.3 |
| what-uncertainties-do-we-need-in-bayesian | 37.3% |
| std2p-rgbd-semantic-segmentation-using-spatio | 40.1% |
| warp-refine-propagation-semi-supervised-auto | 52.2% |
| rfnet-region-aware-fusion-network-for | 48.13% |
| pattern-affinitive-propagation-across-depth-1 | 50.4% |
| spatial-information-guided-adaptive-context | 48.2% |
| depth-aware-cnn-for-rgb-d-segmentation | 43.9% |
| channel-exchanging-networks-for-multimodal | 52.5% |
| dformer-rethinking-rgbd-representation | 51.8% |
| hspformer-hierarchical-spatial-perception | 57.8% |
| hemis-hetero-modal-image-segmentation | 37.77% |
| depth-adapted-cnns-for-rgb-d-semantic | 51.24% |
| real-time-joint-semantic-segmentation-and | 42.0% |
| context-aware-interaction-network-for-rgb-t | 52.6% |
| geminifusion-efficient-pixel-wise-multimodal | 57.7 |
| dformer-rethinking-rgbd-representation | 57.2% |
| temporally-distributed-networks-for-fast | 37.4 |
| contrastive-multimodal-fusion-with | 48.1% |
| cmx-cross-modal-fusion-for-rgb-x-semantic | 56.9% |
| spatial-information-guided-convolution-for | 51.0% |
| cmx-cross-modal-fusion-for-rgb-x-semantic | 56.3% |
| variational-context-deformable-convnets-for | 45.3 |
| fully-convolutional-networks-for-semantic | - |
| understanding-dark-scenes-by-contrasting | 53.7% |
| sosd-net-joint-semantic-object-segmentation | 45.0% |
| efficient-multi-task-rgb-d-scene-analysis-for | 53.34% |
| semantic-segmentation-with-reverse-attention | 41.2% |
| omnivore-a-single-model-for-many-visual | 56.8% |
| attention-based-dual-supervised-decoder-for | 52.5% |
| light-weight-refinenet-for-real-time-semantic | 43.6% |
| shapeconv-shape-aware-convolutional-layer-for | 49.0% |
| cerberus-transformer-joint-semantic | 50.4% |
| comptr-towards-diverse-bi-source-dense | 49.2% |
| hapnet-toward-superior-rgb-thermal-scene | 55.0 |
| shapeconv-shape-aware-convolutional-layer-for | 51.3% |
| depth-adapted-cnns-for-rgb-d-semantic | 49.15% |
| dformer-rethinking-rgbd-representation | 55.6% |
| efficient-rgb-d-semantic-segmentation-for | 50.30 |
| diffusion-based-rgb-d-semantic-segmentation | 61.5 |
| recurrent-scene-parsing-with-perspective | 44.5% |
| cross-task-attention-mechanism-for-dense | 40.84% |
| panopticndt-efficient-and-robust-panoptic | 59.02 |
| depth-adapted-cnns-for-rgb-d-semantic | 47.02% |
| shapeconv-shape-aware-convolutional-layer-for | 48.8% |
| omnivec-learning-robust-representations-with | 60.8 |
| deep-feature-selection-and-fusion-for-rgb-d | 52.0% |
| light-weight-refinenet-for-real-time-semantic | 41.7% |
| geminifusion-efficient-pixel-wise-multimodal | 60.2 |
| prompt-guided-transformer-for-multi-task | 46.43 |
| depth-adapted-cnns-for-rgb-d-semantic | 50.05% |
| dformer-rethinking-rgbd-representation | 53.6% |