Buzz
In the near future, controlling another person’s smile could become a reality. A study presented at the 2015 SIGGRAPH Asia conference by teams from the University of Erlangen-Nuremberg, the Max Planck Institute for Informatics, and Stanford University showcased a breakthrough in real-time facial expression transfer between individuals in video footage [PDF]. This innovation enables my smile to be seamlessly projected onto a CGI version of your face, achieving what the researchers describe as 'unprecedented visual realism.'
A specialized sensor evaluates facial geometry and skin reflectance across various regions. It constructs digital models of both participants' faces and the surrounding lighting, allowing one person’s expression to be overlaid onto another’s. The system calculates the disparity between the source expression and the target face, blending them seamlessly. The aim is to modify only the specific facial areas required for the expression, leaving the rest of the target’s face unchanged.
What practical, non-criminal applications could this have, you might wonder? The researchers propose that this technology could enable remote freelancers to work from home without ever needing to dress formally.
Envision a multilingual video-conferencing system where a participant’s video is dynamically modified to realistically replicate the facial expressions and lip movements of a live translator. Alternatively, picture a scenario where a professionally recorded video of someone in formal attire is reenacted using a real-time facial capture of yourself lounging casually on your couch.
Additionally, the potential for pranks is limitless. Consider the possibility of SNL superimposing Kate McKinnon’s facial expressions onto what appears to be Hillary Clinton’s actual face.
