Figure 3: The unsharp mask is the difference between the test data and a blurred version of that data. In this case, I used a Gaussian Blur with a radius of two, and subtracted the blurred version of Figure 2 from the original.