In this project, I used data taken from a lightfield camera, which captures point/axis specific light information instead of just taking a single picture, and used it to simulate camera effects like depth of field and aperture refocusing.
In this part, I had to choose points on the face so that I could generate a triangulation of features to aid the morph. Pictured above are the points I chose superimposed on my face. Below are the same points and triangles superimposed on Alan Tang Kwong Wing's face.
I generated a changing depth of field effect by shifting pixels based on where the origin point of light was from, multiplied by a certain scalar, and then taking the averages of all the lightfield images processed this way.
Below is pictured the shifts with a scalar of -.1, .2, and .4
Below is all the shifts (50 frames) as a gif.
To simulate aperture shifts as though it were a real camera lens being refocused, simply averaging the images is not enough. By averaging lots of images perpendicular to the optical axis of the focal point, I could simulate this effect.
Below is pictured the refocusing at the least and most weighted:
Below is all the refocusing (9 frames) as a gif.