Images are so unpredictable. They can be too big or too small. If only there was a way to crop images and remove unnecessary background artifacts...
An energy function determines a pixel's importance. Shai Avidan and Ariel Shamir's paper recommends using the sobel gradient function for this application.
To calculate a seam, we sum the lowest important energy path based off our energy function. We must then remove the seams with the lowest energy until we reach the desired dimension.
In this case, seam carving distored Efros' body.
This project's demonstration of dynamic programming in computer vision was extremely insightful. This project connected CS170 algorithms and computer vision to generate amazing imaging techniques.
This project leverages lightfields to refocus images at varying depth and adjust the aperture of the image.
By shifting the image around varying centers and taking the average of all these shifted images, we are able to refocus the image at the specified depth.
To adjust the aperture of an image, we can take the average of a subset of images depending on the aperture size.
This project showcases the advantages of lightfields and surprised me of its ability to refocus images and adjust the image's aperture.