Advanced Computer Graphics, Fall 2014

Written Exercise 2: Computational Photography

Due on Wed, Oct 15 at 11:59PM

Imagine you had a camera that could capture images with 10 million pixels at 1000 frames per second over a duration of several seconds, and that image processing algorithms within the camera were fast enough to analyze each image and select new camera parameters (focal length, aperature, flash intensity, etc.) between each successive image.

1. How would you use such a camera to capture the "best possible" image of your friend standing in front of a tourist site (e.g., Nassau Hall)? Please be specific about how camera parameters would be selected for each image and how images would (possibly) be combined to create the final resulting image.

2. How would such a camera affect representations and storage of digital photographs? Currently, we imagine a digital photograph to be a single image (a 2D array of radiance samples). How will that change when O(10^3) images with O(10^7) pixels are available for each photograph? How and why?

Please write your answers in less than one page total and email the result in pdf format to funk@cs.princeton.edu.