Semantic Segmentation On Pascal Voc 2012
Metriken
Mean IoU
Ergebnisse
Leistungsergebnisse verschiedener Modelle zu diesem Benchmark
Vergleichstabelle
| Modellname | Mean IoU |
|---|---|
| fully-convolutional-networks-for-semantic-1 | 62.2% |
| efficient-yet-deep-convolutional-neural | 69% |
| light-weight-refinenet-for-real-time-semantic | 81.1% |
| light-weight-refinenet-for-real-time-semantic | 82.0% |
| encoder-decoder-with-atrous-separable | 89.0% |
| learning-a-discriminative-feature-network-for | 86.2% |
| squeeze-and-attention-networks-for-semantic | 86.1% |
| espnet-efficient-spatial-pyramid-of-dilated | 63.01% |
| understanding-convolution-for-semantic | 83.1% |
| pyramid-scene-parsing-network | 85.4% |
| parsenet-looking-wider-to-see-better | 69.8% |
| exploiting-saliency-for-object-segmentation | 56.7% |
| fast-exact-and-multi-scale-inference-for | 80.2% |
| waterfall-atrous-spatial-pooling-architecture | 79.6% |
| context-encoding-for-semantic-segmentation | 82.9% |
| self-supervised-difference-detection-for-1 | 65.5 |
| multi-scale-context-aggregation-by-dilated | 67.6% |
| stacked-deconvolutional-network-for-semantic | 86.6% |
| espnetv2-a-light-weight-power-efficient-and | 68.0% |
| the-lovasz-softmax-loss-a-tractable-surrogate | 79.00% |
| reliability-does-matter-an-end-to-end-weakly | 66.5 |
| is-attention-better-than-matrix-decomposition-1 | 85.9% |
| object-contextual-representations-for | 84.3% |
| context-encoding-for-semantic-segmentation | 85.9% |
| auto-deeplab-hierarchical-neural-architecture | 85.6% |
| dual-attention-network-for-scene-segmentation | 82.6% |
| beyond-self-attention-external-attention | 84% |
| refinenet-multi-path-refinement-networks-for | 84.2% |
| simple-does-it-weakly-supervised-instance-and | 72.8% |
| semantic-image-segmentation-with-deep | 71.6% |
| object-contextual-representations-for | 84.5% |
| large-kernel-matters-improve-semantic | 83.6% |
| dicenet-dimension-wise-convolutions-for | 67.31% |
| light-weight-refinenet-for-real-time-semantic | 79.2% |
| triply-supervised-decoder-networks-for-joint | 83.3% |
| the-lovasz-softmax-loss-a-tractable-surrogate | 79.0% |
| multi-path-segmentation-network | 84.2% |
| encoder-decoder-with-atrous-separable | 89.0% |
| conditional-random-fields-as-recurrent-neural-1 | 74.7% |
| boxsup-exploiting-bounding-boxes-to-supervise | 64.6% |
| find-it-if-you-can-end-to-end-adversarial | 63.8% |
| simultaneous-detection-and-segmentation | 51.6% |
| learning-a-discriminative-feature-network-for | 82.7% |
| wider-or-deeper-revisiting-the-resnet-model | 84.9% |
| not-all-pixels-are-equal-difficulty-aware | 82.7% |
| deeplab-semantic-image-segmentation-with-deep | 79.7% |
| pyramid-scene-parsing-network | 82.6% |
| edgenext-efficiently-amalgamated-cnn | 80.2% |
| light-weight-refinenet-for-real-time-semantic | 82.7% |
| squeeze-and-attention-networks-for-semantic | 83.2% |
| rethinking-atrous-convolution-for-semantic | 86.9% |