Semantic Segmentation On Cityscapes
Metriken
Mean IoU (class)
Ergebnisse
Leistungsergebnisse verschiedener Modelle zu diesem Benchmark
Vergleichstabelle
| Modellname | Mean IoU (class) |
|---|---|
| efficient-rgb-d-semantic-segmentation-for | 80.09% |
| esnet-an-efficient-symmetric-network-for-real | 70.7% |
| resnest-split-attention-networks | 83.3% |
| understanding-convolution-for-semantic | 77.6% |
| liteseg-a-novel-lightweight-convnet-for | 70.75% |
| self-supervised-model-adaptation-for | 81.24% |
| object-contextual-representations-for | 83.7% |
| multi-path-segmentation-network | 79.0% |
| rethinking-atrous-convolution-for-semantic | 81.3% |
| internimage-exploring-large-scale-vision | 86.1% |
| squeezenas-fast-neural-architecture-search | 72.5% |
| ocnet-object-context-network-for-scene | 81.7% |
| multi-receptive-field-network-for-semantic | 83.0% |
| liteseg-a-novel-lightweight-convnet-for | 67.81% |
| 190807919 | 81.6% |
| espnetv2-a-light-weight-power-efficient-and | 66.2% |
| liteseg-a-novel-lightweight-convnet-for | 67.81% |
| semantic-image-segmentation-via-deep-parsing | 66.8% |
| 2003-13328 | 82.0% |
| high-resolution-representations-for-labeling | 81.6% |
| lednet-a-lightweight-encoder-decoder-network | 70.6% |
| erfnet-efficient-residual-factorized-convnet | 69.8% |
| denseaspp-for-semantic-segmentation-in-street | 80.6% |
| object-contextual-representations-for | 82.4% |
| the-lovasz-softmax-loss-a-tractable-surrogate | 63.06% |
| sinet-extreme-lightweight-portrait | 66.5% |
| wider-or-deeper-revisiting-the-resnet-model | 78.4% |
| vision-transformer-adapter-for-dense | 85.2% |
| full-resolution-residual-networks-for | 71.8% |
| global-aggregation-then-local-distribution-in | 83.3% |
| waterfall-atrous-spatial-pooling-architecture | 70.5% |
| dynamic-structured-semantic-propagation | 77.8% |
| scene-segmentation-with-dual-relation-aware | 82.9% |
| gff-gated-fully-fusion-for-semantic | 82.3% |
| harnessing-diffusion-models-for-visual | 86.2 |
| semantic-correlation-promoted-shape-variant-1 | 81.0% |
| espnet-efficient-spatial-pyramid-of-dilated | 60.3% |
| learning-a-discriminative-feature-network-for | 79.3% |
| enet-a-deep-neural-network-architecture-for | 58.3% |
| liteseg-a-novel-lightweight-convnet-for | 65.17% |
| fast-scnn-fast-semantic-segmentation-network | 68% |
| ikshana-a-theory-of-human-scene-understanding | 42.07% |
| searching-for-efficient-multi-scale | 82.7% |
| auto-deeplab-hierarchical-neural-architecture | 82.1% |
| semantic-image-segmentation-with-deep | 63.1% |
| learning-a-discriminative-feature-network-for | 80.3% |
| icnet-for-real-time-semantic-segmentation-on | 70.6% |
| dcnas-densely-connected-neural-architecture | 83.6% |
| dual-attention-network-for-scene-segmentation | 81.5% |
| template-based-automatic-search-of-compact | 67.8% |
| efficient-dense-modules-of-asymmetric | 67.3 |
| cars-cant-fly-up-in-the-sky-improving-urban | 83.2% |
| psanet-point-wise-spatial-attention-network | 80.1% |
| fasterseg-searching-for-faster-real-time-1 | 71.5% |
| semantic-aware-generation-for-self-supervised | 76.9 |
| template-based-automatic-search-of-compact | 67.7% |
| squeezenas-fast-neural-architecture-search | 66.8% |
| gated-scnn-gated-shape-cnns-for-semantic | 82.8% |
| pyramid-scene-parsing-network | 78.4% |
| inverseform-a-loss-function-for-structured | 85.6% |
| disentangled-non-local-neural-networks | 83% |
| deep-dual-resolution-networks-for-real-time | 82.4% |
| fully-convolutional-networks-for-semantic | 65.3% |
| regularized-frank-wolfe-for-dense-crfs | 83.6% |
| resolution-aware-design-of-atrous-rates-for | 79.9% |
| ikshana-a-theory-of-human-scene-understanding | 45.02% |
| bisenet-bilateral-segmentation-network-for | 78.9% |
| efficient-piecewise-training-of-deep | 71.6% |
| hs3-learning-with-proper-task-complexity-in | 85.8% |
| incorporating-luminance-depth-and-color | 71.3 |
| vltseg-simple-transfer-of-clip-based-vision | 86.4 |
| self-supervised-model-adaptation-for | 82.3% |
| refinenet-multi-path-refinement-networks-for | 73.6% |
| dual-graph-convolutional-network-for-semantic | 82% |
| object-contextual-representations-for | 83.0% |
| in-defense-of-pre-trained-imagenet | 75.5% |
| boundary-aware-feature-propagation-for-scene | 81.4% |
| recurrent-scene-parsing-with-perspective | 78.2% |
| k-means-mask-transformer | 83.2% |
| object-contextual-representations-for | 81.8% |
| efficientps-efficient-panoptic-segmentation | 84.21% |
| depth-anything-unleashing-the-power-of-large | 84.8% |
| sernet-former-semantic-segmentation-by | 84.83 |
| swinmtl-a-shared-architecture-for | 76.41% |
| rethinking-semantic-segmentation-from-a | 81.64% |
| semantic-segmentation-with-multi-scale | 72.4% |
| pyramid-scene-parsing-network | 80.2% |
| panoptic-deeplab-a-simple-strong-and-fast | 84.2% |
| searching-for-mobilenetv3 | 72.6% |
| asymmetric-non-local-neural-networks-for | 81.3% |
| ikshana-a-theory-of-human-scene-understanding | 54.82% |
| joint-semantic-segmentation-and-boundary | 81.8 |
| laplacian-pyramid-reconstruction-and | 71.8% |
| densely-connected-multidilated-convolutional | 80.8% |
| multi-scale-context-aggregation-by-dilated | 67.1% |
| liteseg-a-novel-lightweight-convnet-for | 65.17% |
| deeplab-semantic-image-segmentation-with-deep | 70.4% |
| segnet-a-deep-convolutional-encoder-decoder | 57.0% |
| dfanet-deep-feature-aggregation-for-real-time | 71.3% |
| ccnet-criss-cross-attention-for-semantic | 81.4% |
| object-contextual-representations-for | 84.5% |
| caa-channelized-axial-attention-for-semantic | 82.6% |
| context-prior-for-scene-segmentation | 81.3% |
| segformer-simple-and-efficient-design-for | 83.1% |
| adaptive-affinity-fields-for-semantic | 79.1% |