ABSTRACT
Peking opera is one of the important representatives of Chinese traditional culture. Especially, its unique facial art has high cultural value, research value, and application value. Scale Invariant Feature Transform (SIFT) is a scale-invariant feature transformation algorithm proposed by David G. Lowe in 1999 and refined in 2004. This chapter describes the SIFT feature the facial images of Peking opera. The SIFT feature vectors are extracted and dimensions are reduced and normalized for illumination factors. Support Vector Machine (SVM) classifiers are used to classify the facial images of Peking opera. SIFT uses the gradient direction feature of the neighborhood pixels of the feature points to determine the direction of the feature points. Among the methods of classification and recognition, Cortes and Vapnik proposed the SVM algorithm as an effective classifier, which shows many advantages in solving small sample, nonlinear, and high-dimensional pattern recognition.
