Rephotography, the ability to take a photograph from the same perspective at two different times, is a synergy of techniques from computer vision and computer graphics. The ability to take such pictures gives a user the opportunity to make photo composites with interesting temporal features, for example a picture with the same person in multiple poses, or a blended photo of a historical photo and a modern photo of the same scene. My work will be to implement a high-accuracy rephotography system for iOS with a polished user interface using algorithmic techniques from both computer vision and computer graphics.

I have worked with Professor Tomaso Poggio on a project involving action recognition in videos and a separate project about object recognition models from videos. Last year I spent some time building a photo-recognition system to recognize and size keys, key blanks, and key cuts for ease of use in automated key copying. This past summer, I worked in at a laboratory at the Universidad Politecnica de Madrid developing an algorithm for more effective image mosaicing from a video captured by a flying UAV.