Confusion about density calculation

[daisyranc]

Thank you for your excellent work!!

After reading your paper and code, I'm confused about density calculation. I have tested on my own data. I found that the density calculated by your code may be different from the density I observed.

Here are the details of my test experiments:

I randomly sample one point cloud (bunny) from 200000 to 100 (this is the sample number list: [200000,100000,50000,10000,5000,1000,500,100]), and calculate the density for each sampled point cloud. I think the density will be smaller and smaller as the point numbers become smaller and smaller. However, here are my results:

point number:200000
density:0.40053454
point number:100000
density:0.40531346
point number:10000
density:0.39948502
point number:5000
density:0.4131956
point number:1000
density:0.41057622
point number:500
density:0.4307725
point number:100
density:0.4311616

Could you give more explanation about the KDE you used in your paper? Thank you so much!!

[daisyranc]

I used the mean value of all points' densities to represent point cloud density here.