Nicolas Mellado
Hi, Could you please share the point cloud so I can understand what is going wrong ?
Ok found it: your files do not have normals, so the GLS computation is wrong. I got good results by using the filter _computing normals for point clouds_ in meshlab,...
Here you go: I used the default parameters for the filter, and it worked well.
There is currently a bug with file saving, see https://github.com/STORM-IRIT/pcednet-supp/issues/11.
> There is a bug occurring during the loading of some PLY files which not contain mesh information. > > Some files with only vertex information (with normal) given work...
Did we had this bug before your PR ?
Bonjour @lblre , (désolé pour le temps de traitement) Est ce que le problème est toujours présent avec la dernière version ?
This version brings a lot of deprecation warning that need to be fixed. I think we can keep it to Milestone 2.0 as already planned.
Could you implement the CNC algorithm directly ? https://github.com/dcoeurjo/CorrectedNormalCurrent
@Me-k-01 I think we can start with this PR 👍