The first part of this project involved refocusing images using images taken from a light field camera. To do this, I shifted the images by how much their uv positions differed from the central position of 8, 8 scaled by a constant c. Some examples of different shifts are listed below.
The second part of the project involved modifying the aperture size of the camera. To do this, I defined a radius, then used all the pictures within a radius r square of the center to create refocus the images.
For this part, I put together a texture with randomly selected patches and no other processing
For this part, I put together a texture that calculated the ssd between overlapping patches to find the best match for every patch.
For this part, I did the same thing as overlapping patches but stiched together the seams to create a better image.
For this part, I added an additional cost to the ssd that was the difference between the reference image and the sample image.
