The Research On Points Distribution In High-Dimensional Spcace Geometry For Biomimetic Pattern Recognition

In digital era, images, voice etc. all can be transformed to multi-dimensional numbers of high-dimensional space. The development of high-dimensional geometrical methods rises a new way on finding new directions of the information science.This paper aims at researching on points distribution in high-dimensional space geometry for biomimetic pattern recognition. In this paper, the basic conceptions and analysis method of high-dimensional geometry are presented, and we also analyse the properties of the points distribution in high-dimensional space, several classical computational issues are discussed subsequently. As for the application of this theory, we research its application on image restoration and face recognition.We have studied the points distribution of images which are mapped to high-dimensional space then propose a novel image restoration approach based on high-dimensional space geometry. Begin with the original blurred image, we get two further blurred images, then the restored image can be obtained through the regressive curve derived from the three points which is mapped form the images. Experiments have proved the availability of this "blurred-blurred-restored" algorithm.For the limitation of the traditional statistical pattern recognition approaches which set optimal separating as its main principle, and considering the superiority of the biomimetic recognition, we propose an recognition approach based points distribution in high-dimensional space geometry. The human images' distribution in high-dimensional space is studied. Using the multi-weight neuron neural networks which are based on the method of covering the high dimensional geometrical distribution of the sample set in the feature Space, experiments have proved the availability of this approach.