Author: Mikhail Skobov

Below are the explanations for the questions that required them:

2. These images look the way they do because of the edge detection function used. It detected
the amount of change between neighboring pixels and gave them an amplitude. This was then mapped
to a color scale with red being high energy, and blue low energy. It is easy to see that distinct
shapes have higher energy from this computation.

3. The seams (shown in red) are optimal because that is where the minimal change occurs between
neighboring pixels. If we take a look at the images from part 2, we can see that the seams occur
in mostly blue regions, hence have low energy, and are good first seams.

4. The grassy knoll had a very successful result. This is most likely because the algorithm was able to find the lowest energy seams in the sky, rather than trying to work its way around the blades of grass/branches.

5. This image had several flaws which made it a poor choice for size reduction. First, there 
were not very many straight lines, which most likely made it hard to distinguish easily between 
the best possible seams. Also, the colors were not very contrasting, so peoples faces got removed.
Lastly, there was too much going on in the whole image, which is also bad for finding the best seam.