We present a method to reduce the computational complexity of the depth from defocus problem with an uncalibrated
handheld camera for the purpose of synthetic defocus. This is done by using a bilateral-space approximation of the images
from the camera. The method shows a 50-200x computation time reduction for the bottlenecked steps within the state-ofthe-
art method and shows similar reduction for the total time as input size increases. The results show a similar defocus
render between the proposed method and the state-of-the-art.
handheld camera for the purpose of synthetic defocus. This is done by using a bilateral-space approximation of the images
from the camera. The method shows a 50-200x computation time reduction for the bottlenecked steps within the state-ofthe-
art method and shows similar reduction for the total time as input size increases. The results show a similar defocus
render between the proposed method and the state-of-the-art.