Synthesizing Obama
The next frontier in fakery, 2017
Supasorn Suwajanakorn, Steven M. Seitz & Ira Kemelmacher-Shlizerman (US)
This exhibit showcases a visual form of lip-syncing, with a neural network trained on many hours of past footage used to convert audio files of an individual’s speech into realistic mouth shapes, which are then grafted onto and blended with the head of that person from another existing video. A team of researchers from the University of Washington used the process to successfully generate a highly realistic video of former president Barack Obama talking about terrorism, fatherhood, job creation and other topics, using audio clips of those speeches and existing weekly video addresses that were originally on a different topic.
BIO
Supasorn Suwajanakorn, Steven M. Seitz, Ira Kemelmacher-Shlizerman are researchers at the University of Washington. For more information on their project, you can visit their project page at their website here.