The number of points

[rsy6318]

Hi Wanquan,
In PUGeo-Net, during the testing, the number of points is 5000 and that in NeuralPoints is 2000. Would it influence the result?

[WanquanF]

We should just keep the same number of input points for all comparing methods in the experiments.

[kangya998]

We should just keep the same number of input points for all comparing methods in the experiments.

Hi，I found that the test points in the paper is set to 10000, but the code is set to 2000, why is that? How can the final merged generated point cloud be evaluated with the ground truth (40000, 160000) point cloud, and finally I found that the merged generated point cloud size is different from the mesh size of the ground truth, am I missing something?