mrakotosaon/pointcleannet

PointCleanNet does not export "normal" and "xyz" file?

linlinge opened this issue · 5 comments

linlinge commented

When I run PointCleanNet with my own dataset, it just export ".info" file but missing the "normal" and "xyz" file.

mrakotosaon commented

Hi @linlinge ,

For outliers removal the output is a file .info that contains the xyz coordinates in the first 3 columns and the outlier label in the 4th column. We decide that a point is an outlier if that value is larger than 0.5. We do not regenerate the xyz file for the task of outlier removal.

linlinge commented

Thanks for your reply! I will try it again with your recommand threshold.

Hao-HUST commented

Hi @linlinge ,

For outliers removal the output is a file .info that contains the xyz coordinates in the first 3 columns and the outlier label in the 4th column. We decide that a point is an outlier if that value is larger than 0.5. We do not regenerate the xyz file for the task of outlier removal.

Hi, @mrakotosaon.

How to visualize the outlier remove result(the .info file)?

Hao-HUST commented

Hi @linlinge ,
For outliers removal the output is a file .info that contains the xyz coordinates in the first 3 columns and the outlier label in the 4th column. We decide that a point is an outlier if that value is larger than 0.5. We do not regenerate the xyz file for the task of outlier removal.

Hi, @mrakotosaon.

How to visualize the outlier remove result(the .info file)?

Oh, the .info file(outlier result) can be visualized by the software of cloud compare, and the outlier labels can be treated as the scalar field value.

It is easy to separate the outlier and inlier point cloud by setting a scalar threshold.

mrakotosaon commented

Hi @Hao-HUST
That's nice to hear! :)
Another possibility would also be to filter the file with a small python script if you want a permanently filtered file.
Best,