Deblur can be implemented in four different algorithms. They differ by computation and memory demands. Deblurring methods generally amplify digital noise in photo. The first implemented method can work with noise and does not amplify it as much. The second method is fast, and does not need much memory. The third method deblurs very aggressively and the fourth method is absolutely stable and does not diverge anytime.
What's new in this version:
Version 2.2 has transparency supported.