Comprehensive Evaluation of Feature Extraction Method with Multi-measurement

摘要

This paper focus on the evaluation of several different feature extraction methods in pattern recognition. It provides three kinds of measurement including minimum distance,sum of distance,variance of distance between classes,and adopt AHP (Analytic Hierarchy Process) method to evaluate the performance of features which were extracted from thirty-six characters by different feature extraction method. This paper also gives quantitative evaluations result of feature extraction methods with dimension 32,64 and 96 respectively.