2D Semantic Segmentation on KITTI-360 (val)
68.1mIoUDeepLabV3+
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| DeepLabV3+Backbone=Xception-65, Input=RGB (fixed imaging), Params (M)=~422025.12 | 68.1 | — | |
| Joint Optics-Sensor-Model Co-DesignInput=RAW-to-task (optics + sensor + model), Params (M)=0.47, Epochs=502025.12 | 67.21 | 91.83 | |
| DeepLabV3Backbone=ResNet-101, Input=RGB (fixed imaging), Params (M)=~602025.12 | 66.4 | — | |
| PSPNetBackbone=ResNet-101, Input=RGB (fixed imaging), Params (M)=~652025.12 | 64.9 | — | |
| FCN-8sBackbone=VGG-16, Input=RGB (fixed imaging), Params (M)=~1342025.12 | 54 | — |