Ear Recognition Based on 2DLDA and FSVM
Han Jian · Science Technology and Engineering · 2012
The University enterprise R SeD Center of Measuring and Testing According to the existing problems on feature extraction and recognition of the human ear image,a recognition method of the human ear image is put forward,which combins the 2-dimensional linear discriminant analysis(2DLDA) and fuzzy support vector machines(FSVM).2DLDA that directly projects to extract ear features can keep a large number of information of the human ear image on within-class and between-class samples.Meanwhile,FSVM based on the support vector machines(SVM) introductes membership parameters,which are more suitable for multi-class problems.The experimental results show that the method has higher recognition rate than 2DLDA.