Mesh vertex vs. point cloud position

If one were to compute a non-decimated reconstruction using laser scan files only , would the output meshes vertices be the exact same as the input point cloud on a point by point basis?

 

In other words, does RC modify input points, create new points or delete points from input point clouds.  I seem to recall that the answer is no, but I can’t find the post and it was from years ago.

 

Thanks