Supasorn Suwajanakorn, Ira Kemelmacher-Shlizerman, Steven M. Seitz target videos taken under uncontrolled and uncalibrated imaging conditions, such as youtube videos of celebrities. In the heart of this work is a new dense 3D flow estimation method coupled with shape from shading. Unlike related works scientisits do not assume availability of a blend shape model, nor require the person to participate in a training/capturing process. Instead they leverage the large amounts of photos that are available per individual in personal or internet photo collections.
Below you can see the results for a variety of video sequences that include various lighting conditions, head poses, and facial expressions. This video was made for European Conference on Computer Vision (ECCV 2014) which is held on September 6, 2014 in Zurich (Switzerland).
More about tis topic here.
LEAVE A COMMENT
COMMENTS