The difference between PTv3 and OACNNs
jing-zhao9 opened this issue · comments
PTv3 and OACNNs both use 3D sparse convolution and do not use KNN, so can we assume that they are both voxel based methods?
OACNNs are voxel-based methods, yet PTv3 is a serialization-based method, as defined in our paper. As point cloud in PTv3 is structured and ordered.