Semantic Occupancy Grid Mapping on Cityscapes weak ground truth (val)
91.8Mean Accuracywith disparity
Evaluation results
| Method | Links | ||
|---|---|---|---|
| with disparityInput Mode=Binocular, Feature Extraction=VGG-16 based FCN [1]2018.04 | 91.8 | 80 | |
| VEDArchitecture=Variational Encoder-Decoder2018.04 | 73.8 | 59.5 | |
| SegNetArchitecture=CNN Encoder-Decoder, Backbone=VGG-16 [12]2018.04 | 73.7 | 61.1 | |
| flat-plane assumptionInput Mode=Monocular, Feature Extraction=VGG-16 based FCN [1]2018.04 | 69.9 | 47.2 |