A New Method Based on Wavelet Packet Transformation in 3D Face Feature Extraction

Abstract

Abstract.Transition from 2D to 3D in face pattern recognition in order to seek richer human face information but feature extraction as an important step among them can be seen as information compression. Obviously the two are contrary. For the fast effective locate and extract features in the large-scale 3D face data set, the highly effective digital compression computation becomes very important. This paper study a feature extraction method based on the wavelet packet transformation compression, proposed the specific method and the principl.When the actual data which distributed in 2D or higher dimensional projection space disobeyed assuming require or sample parameters inaccurate estimated ,it cannot be correctly classified drawbacks or be correctly classified. Our experiment is indicated that our method can not only conquer this disadvantage but also meet the actual demand at the processing time. The experiment is indicated that our method towards3D face feature extraction data sets compress processing is effective and feasible.