Computer Vision News - May 2016

COMPUTER VISION NEWS 19 Face2Face: the result Research Free Subscription Dear reader, Do you enjoy reading Computer Vision News? Would you like to receive it for free in your mailbox every month? It will take you less than 1 minute to fill the Subscription Form. Join many others computer vision professionals and receive all the next issues of Computer Vision News as soon as they are published. BTW, we hate SPAM and promise to keep your email address safe, always . You can also read Computer Vision News on our website and find in our archive new and old issues as well. Subscription Form (click here, it’s free) The end result of this model is at least two-fold: on one hand it enables the real time animation of legacy video footage (from YouTube or any other video hosting platform), which might have an influence in the entertainment field. On the other hand, it might open new ways for teleconferencing, Virtual Reality and Advanced Reality applications . Even some videos with translated audios might be dubbed with surprising results. And finally, from now on, beware when you watch a video of Donald Trump speaking: they might be his real lips but somebody else’s mouth movements! The authors are careful to list several assumption and limitation of the presented method: (1) assumption of Lambertian surfaces and smooth illumination is limiting; (2) face occlusions by long hair and a beard might not work well; (3) the mouth synthesis assumes sufficient visible expression variation in the target sequence. However, on a too short sequence or when the target remains static, the method will fail to learn the person-specific mouth behavior.

RkJQdWJsaXNoZXIy NTc3NzU=