This project seeks to explore depth refocusing and aperture adjustment using a series of images from the stanford light field archive.
1: Depth Refocusing
The following are images and gif are based on depth refocusing. It is possible to use the 17x17 grid of the cameras from the Stanford light field
to sharpen different aspects of the scene, functionally adjusting the depth at which the image is focused on.
This gif shows the product of focusing on different images based on using different offset values.
2. Aperature adjustment
The following gif shows the ability to adjust our aperature by averaging subsets of the image based on how far they are from
from the center of our grid. The end of the grid shows choosing a very large offset, causing extreme blur.
Incorrectly Detected.
I learned a lot about how artificially changing aperature and depth of focus in images is performed with the correct dataset.
I was amazed at how relatively simple the code was, and really enjoyed the process of creating this project.
Project 2: Image Quilting.
This project focuses on the image quilting algorithm for texture synthesis and transfer. It is broken into a few sections,
but only randomly sampled texture and overlapping patches were implemented.
1. Randomly sampled texture
This was functionally done by determining a size of patches of the image taken, an output size for the output, and uses the data
to cut random samples of the pattern in the image and use continually paste the sample into the output.
2. Overlapping patches
This function takes sample patches out of the image, and compares the overlap cost if it was lined up with another patch.
If the SSD between the patches is low enough, the patches are pasted together for the output. The goal is to repeat the pattern seamlessly.
I had a lot of fun with this project, but sadly I contracted covid right before dead week. This made me fall extremely
far behind in coursework so I was not able to complete it. I really enjoyed this class!