A Survey of Deformable Human Face Model
Xiaoyang Tan · Journal of Shenyang Aerospace University · 2011
Tracking and understanding face images is the first and one of the key steps to construct any intelligent system which works automatically in complex environment with human beings involved.One way to achieve it is to build flexible and compact deformable model for object using statistical learning technique.In this paper,we survey the recent advances on how to build a highly accurate and robust deformable model for human face images under difficult conditions.In particular,under a uniformed Bayesian framework,we describe and analyze the prominent algorithms and provide some insights into the underlying principles,interactions,advantages,limitations and design tradeoffs of these techniques.Several promising directions for future research are also proposed in this paper.