Add the semantic category of blind path during training

[SZUshenyan]

Hello, I want to add the semantic category of blind path during training. May I ask if I can only collect image data and voxel the image data into the format of semantic kitti dataset?