Semantic Segmentation on Sift Flow
86.8Pixel AccuracyMultiscale Convolutional Network
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Multiscale Convolutional NetworkBackbone=VGG, Class Reweighting=false2014.11 | 86.8 | 46.4 | 77.9 | 38.8 | |
| Long et al.2014.11 | 85.1 | 51.7 | 76.1 | 39.5 | |
| Multiscale Convolutional NetworkBackbone=AlexNet, Class Reweighting=false2014.11 | 84 | 42 | 73.7 | 33.1 | |
| Multiscale Convolutional NetworkBackbone=VGG, Class Reweighting=true2014.11 | 83.8 | 55.7 | 74.7 | 37.6 | |
| Multiscale Convolutional NetworkBackbone=AlexNet, Class Reweighting=true2014.11 | 81.6 | 48.2 | 71.3 | 32.6 | |
| Tighe et al.2014.11 | 78.6 | 39.2 | — | — | |
| Farabet et al.Class Reweighting=false, Note=empirical distribution2014.11 | 78.5 | 29.6 | — | — | |
| Pinheiro et al.2014.11 | 77.7 | 29.8 | — | — | |
| Farabet et al.Class Reweighting=true, Note=rebalanced distribution2014.11 | 74.2 | 46 | — | — |