About number of points

[soon0698]

Hi, thank you for your work.

I noticed that the number of points in this method is almost 2-3 times higher for the modified densification process (e.g., bicycle 9-10M).

Have you conducted any analysis with the same number of points? For instance, other densification methods like AbsGS and FreGS compared the original 3DGS with modified gradient threshold to match point counts.

I just wonder about your method's effectiveness in a fair comparison setting.

Thanks.