Would be nice if I could give it an initial transformation

[shrinivas96]

Somehow the algorithm thinks it has aligned the two scans, but when I plot them they are far from aligned, I wonder if there is a way to input an initial transformation that the algorithm could work from. The function arguments do not seem to have such an argument.

From how much I understand ICP (which is very little), it still needs an initial estimate. But your algorithm does not seem to need that. Though I did not peek inside the hood too much, so i am not sure if this was a trivial thing on my side.