作者: T.F Cootes , G.V Wheeler , K.N Walker , C.J Taylor
DOI: 10.1016/S0262-8856(02)00055-0
关键词:
摘要: Abstract We demonstrate that a small number of 2D linear statistical models are sufficient to capture the shape and appearance face from wide range viewpoints. Such can be used estimate head orientation track faces through large angles. Given multiple images same we learn coupled model describing relationship between frontal profile face. This predict new views seen one view constrain search algorithms which seek locate in simultaneously.