Face Recognition Based on Biomimetic Pattern Recognition by High-Dimensional Geometry

High-dimensional geometry is a vehicle to achieve biomimetic pattern recognition, which is a new model for recognition science and always to construct convex cell bodies for covering samples points in the space. By this way, in this paper, some k- dimension simplex were constructed for the purpose of covering the sample points of each class in the feature space, and then computed each the shortest distance between a sample waiting for recognizing and each simplex for face recognition. The method was applied for some famous face database, and the experimental results indicated that it was better than that of SVM,and indicated its effectiveness and feasibility.