Scalable 3D Panoptic Segmentation As Superpoint Graph Clustering

Benchmark Model Rank Results
3d-semantic-segmentation-on-dalesSuperCluster#4mIoU: 77.3Model size: 210M
3d-semantic-segmentation-on-kitti-360SuperCluster#8miou Val: 62.1Model size: 790K
panoptic-segmentation-on-s3dis-area5SuperCluster#1PQ: 50.1RQ: 60.1SQ: 76.6PQ (with stuff): 58.4
panoptic-segmentation-on-scannetSuperCluster#2PQ: 58.7PQ_th: 69.1PQ_st: 84.1
panoptic-segmentation-on-scannetv2SuperCluster#3PQ: 58.7RQ: 69.1SQ: 84.1Params (M): 1
semantic-segmentation-on-s3disSuperCluster#11Mean IoU: 75.3Number of params: 0.21MParams (M): 0.21
semantic-segmentation-on-s3dis-area5SuperCluster#31mIoU: 68.1Number of params: 0.21