In your method you downsampled the LLFF dataset by a factor of 8, thus, you have divided the projected pixels of each 3d points in the point clouds by the same factor. Will it means when the rays are shooter through this newly scaled pixels, it will meet the same exact 3d points of the point clouds before scaling?
Does this scaling by factor 8 mean scaling the entire scene in the world coordinate system by 8 in x,y and z direction?
I am using surface normals extraction, so I want to know whether this scaling would change my direction of the surface normals.
In your method you downsampled the LLFF dataset by a factor of 8, thus, you have divided the projected pixels of each 3d points in the point clouds by the same factor. Will it means when the rays are shooter through this newly scaled pixels, it will meet the same exact 3d points of the point clouds before scaling?
Does this scaling by factor 8 mean scaling the entire scene in the world coordinate system by 8 in x,y and z direction?
I am using surface normals extraction, so I want to know whether this scaling would change my direction of the surface normals.