Global Feature-based Image Classification and Recognition In Small Sample Size Problem

摘要

In this paper, we designs an image classification and recognition systems. With a series approaches of image pre-processing, segmentation and tracking, extract the global image characteristic parameters (the characteristics of the whole picture). According to obtain the characteristic parameters to realize the image classification and identification. Since this article mainly related to small sample data, so the use of nuclear multi-feature linear discriminant analysis (kernel Multi-feature FLDA, short kMFLDA), KNN method and support vector machine (SVM) Comparison of three methods of classification and recognition rate, experimental results show that the Support Vector Machine (SVM) classification recognition rate higher, more reliable.