PointCNN: Convolution On $\mathcal{X}$-Transformed Points

Benchmark Model Rank Results
3d-instance-segmentation-on-s3disPointCNN#16mIoU: 65.39%mAcc: 75.61
3d-part-segmentation-on-intraPointCNN#4IoU (V): 93.59IoU (A): 74.11DSC (V): 96.62DSC (A): 81.74
3d-part-segmentation-on-shapenet-partPointCNN#28Instance Average IoU: 86.14Class Average IoU: 84.6
3d-point-cloud-classification-on-scanobjectnnPointCNN#67Overall Accuracy: 78.5Mean Accuracy: 75.1OBJ-BG (OA): 86.1
few-shot-3d-point-cloud-classification-on-1PointCNN#24Overall Accuracy: 65.41Standard Deviation: 8.9