What if we have to handle both translation and rotation?
	• A single pattern point p and single scene point s aren’t enough to determine a transformation • Instead, we need 2 pattern points and 2 scene point – The first (p,s) pair gives the translation; the second gives the rotation – This leads to an algorithm that runs in expected time O(m3n2) • If more-complicated transformations are allowed then more points are needed and the time bound grows – Example: For 3D translation and rotation, three points are needed and the time bound is O(m4n3)
			CS409 - Spring 2000			10