Feature Extraction for Integrated Pattern Recognition Systems

Conventional pattern recognition systems have two components: feature analysis and pattern classification. Feature analysis is achieved in two steps: parameter extraction and feature extraction. Feature extraction and pattern classification can be conducted independently or jointly. Two popular independent feature extraction algorithms are Linear Discriminant Analysis (LDA) and Principal Component Analysis (PCA). Minimum Classification Error (MCE) algorithm and Support Vector Machine (SVM) are two integrated pattern classification algorithms. This paper investigates the two integrated pattern classification algorithms. A generalized structure is proposed to enhance the performance of MCE and SVM.