Trigonometric Based on Geometric Algebra Representation Principle for Pattern Recognition

Geometric algebra has been successfully extended to signal processing, feature extraction and pattern recognition, for example Clifford SVM, Clifford neural network, Clifford Fourier transform. But the two-dimensional visual pattern recognition research based on geometric algebra representation principle is still limited. We proposed a theory based on geometric algebra that generates the class field using the triangular method. The values of the geometric product by triangle geometry determine the types of samples tested. The test results of IRIS data set prove our idea, the recognition rate of 98%.