Abstract:

The conversion of a special single sample problem into a general pattern recognition problem is studied by generating several different virtual images using easy 2D transformation in the scope of the plane.The 3D space face pose problems are discussed by mapping 3D movement in 2D plane to generate vivid multi-pose face image.The shortages of the existing methods are analyzed and a modified asynchronous rotational mapping method on the basis of the original method is presented.The experimental results show that the generated images are more rotatable and realistic with lower distortion rate and effectively improved recognition rate.

Key words: single sample;multi-pose;asynchronous rotation;face recognition