FIDNet: LiDAR Point Cloud Semantic Segmentation with Fully Interpolation Decoding

Benchmark Model Rank Results
robust-3d-semantic-segmentation-on-nuscenes-cFIDNet#12mean Corruption Error (mCE): 122.42%
robust-3d-semantic-segmentation-on-robo3dFIDNet (64x2048)#15mean Corruption Error (mCE): 113.81%
semi-supervised-semantic-segmentation-on-23Sup.-only (Range View)#9mIoU (1% Labels): 33.1mIoU (10% Labels): 47.7mIoU (20% Labels): 49.9
semi-supervised-semantic-segmentation-on-25Sup.-only (Range View)#11mIoU (1% Labels): 38.3mIoU (10% Labels): 57.5mIoU (20% Labels): 62.7