When generating ring_map and ray_map, I don't understand why ring_map and ray_map will be accumulated and normalized, can you please help me to understand these two code?
This is a trivial modulation that makes transformation using matrixvt more likely to VoxelPooling. However it almost wont affect final performance (i,e, using binary matrices is also fine)
Hi,
When generating ring_map and ray_map, I don't understand why ring_map and ray_map will be accumulated and normalized, can you please help me to understand these two code?