colorize_naive.m

Here are the results from running the naive algorithm on the smaller images:

Image	AlignmentG	AlignmentR
	[1 -2]	[1 -3]
	[-1 0]	[-6 1]
	[0 2]	[5 3]
	[-3 -1]	[-2 -2]
	[8 1]	[15 1]
	[6 2]	[14 4]
	[5 0]	[10 0]
	[5 1]	[0 0]
	[5 0]	[11 -5]
	[6 2]	[13 3]
	[6 3]	[13 4]
	[8 0]	[13 0]

colorize_pyramid

Here are the results from running the pyramid algorithm on the large images:

	[10 1]	[80 32]
	[136 31]	[120 52]
	[44 -17]	[92 -33]
	[52 15]	[112 40]
	[52 32]	[96 29]
	[72 29]	[140 31]

Bells and Whistles

For the bells and whistles, I cropped the images to remove as much of the border with the misaligned plates as possible.

I did this by finding the corner of the plate that was pushed the lowest and to the left and the diagonal corner that was the highest and most right.

I then cropped the image based on these points