What is the role of coarse global Gaussian priority in the final partitioning results? #171
Description
For city v1, my question is:
Did all point clouds in the scene participate in the training of each block? Is coarse global Gaussian priority just to train better at the boundaries of each block? To facilitate direct concatenation of different blocks without processing (from def blockMerge())?
How to obtain each block in the end?
对于citygs v1，我的问题是：
场景的全部点云参与了每个分块的训练？coarse global Gaussian priority只是让每个分块的边界处训练的更好？以方便不同分块不做处理然后直接拼接（来自def blockMerge()）？
最后如何得到每个分块？
thank you