ZhuLifa

Performance on Semantic3D and S3DIS

4

## Semantic3D I submitted Semantic3D results and got the following results: - semantic-8: 0.690 - reduced-8: 0.758 ## S3DIS I got the following results and much lower than the paper...

Different Backbone between Repo Code and Architecture Paper ?

2

Thanks for your great work and the open-source code. When i read the code, i find one thing: the RPN Network implemented in https://github.com/lijx10/USIP/blob/master/models/networks.py is **different** from the architecture described...

Implemention and EMD Loss

2

Hi vinits: Thanks for your nice work and open source code. I find the following code is not good, because the normalization operation almost **removes the translation** between the source...

Hi @zhudelong, thanks for your great work VDB-EDT. I have some question regarding certain points in the paper. 1. `the lower function is with higher efficiency than the raise function`...