Hanyin Cheng
Computational Photography
Seam Insert, Amplication, and Object Removal/ Preservation

This is the original city image This is the city image after inserting 150 pixels. Seam insertion works similar to seam removal because it uses the same algorithm, but instead of deleting it adds those seams. You can see that this image is slight larger than the original. However it adds the seams in a smart way because it finds the a seam with the least gradient.
This is the original bike image This is the bike image after amplication. Amplication works by first resizing the image using the traditional imresize function, but then uses the seam removal function to remove the "unnecessary" seams, preserving the important content thus giving it an amplication effect. The bike looks bigger than the original, while preserving its dimensions.
This is the bike image after trying to preserve the "seat" on the bike. As you can see, the function attempted to delete all seams except the seat, which looks unedited. The preservation function works similarly to the deletion function excet the region is multiplied with a large gradient multiplier instead of small and negative one This is the bike image after removing the "seat" on the bike. As you can see the image is bike itself is preserved except for the seat As discussed before, this is done by using the roipoly function which helps us select a region or object to add a negative gradient multiplier by. When the seamremoval function loops through, it finds the seams that represent the seat, which are highlighted by the gradient multiplier, and cuts it out.