Perception based 3D face modeling and its applications
Diao Xian-feng · Journal of Chongqing University of Posts and Telecommunications · 2009
Human beings are born with a capacity of restoring the 3D shape from a single planar image.It's an essential research topic in perceptual computing to develop novel approaches for modeling 3D human face from a 2D image via simulating this perceptual procedure.Using the statistic theory,this paper proposes a two-step face modeling scheme based on a facial shape statistic model,which serves as the prior knowledge to constrain the following modeling steps.Realistic 3D faces could be generated with this method.Good results have achieved by applying 3D faces to the problem of illumination and pose estimation.