What happens if I try to predict from 16 channel to 128 channel ?

[ArghyaChatterjee]

Hey I was wondering if I try to increase the resolution from 16 to 128 channels ? In your work, you mentioned you can do from 16 up to 64 but can that go up to 128 ?

[Petros626]

have you been you succesful with this approach?

I guess you should retrain the model with e.g. the KITTI dataset (64 layers) and then try to achieve realistic 128 layer point clouds.

here is the setting of the data you want to achieve:

lidar_super_resolution/scripts/data.py

Line 10 in 37bcccc

data_set = 'carla_ouster_range_image' # high-res training dataset, obtained from simulation - CARLA

I also recomment reading their paper related to some parameters, which could create even better results. You could also try methods like LiDAR4D or NeRF-LiDAR.