Using a series of chessboard images freely provided by the Stanford Light Field Archive, we simulated the effect of digital refocusing. To do so, we modified the original displacements, which are given in the image file names, by finding the relative displacements to the center image, and then multiplying this relative displacement by some constant factor.
Below are the final images as the constant ranges from 0.0 to 0.5.
Here is the smoothly-transitioning GIF of the results:
Using the same images as the previous section, but with a slightly different technique, we can simulate post-capture aperature adjustment. Instead of modifying the constant factor, thereby modifying the focus point, we instead keep the constant the same (in the examples below, at zero), but instead choose subsets of the images around the center image to use for averaging. Increasing the number of images used simulated the effect of a larger aperature.
Below are the final images as the number of images used ranges from 1 to 225.
Here is the smoothly-transitioning GIF of the results: