Georeferencing and maintaining original image dimensions

[phyauna]

Hello,
I have been working with your algorithm on IGN's HD LiDAR data for some time now. I'm encountering difficulties with the output rasters.

Problem Description :
The output raster dimensions change from the original input
Final outputs are consistently resized to 128x128 pixels
This results in reduced quality compared to the input data
I cannot find a way to properly georeference the output rasters

Questions :

• Is there a method to georeference the output rasters while preserving their spatial accuracy?
• Is it possible to maintain the original input image format and dimensions instead of the default 128x128 pixel output?

Thank you for this excellent work. Any guidance on these issues would be greatly appreciated.
Best regards