Classification of traditional Chinese paintings based on supervised learning methods

The potential and feasibility of applying the knowledge of supervised learning methods to Chinese traditional painting classification is discussed and evaluated. Data bases of different artists and categories are collected, from which numerical features are extracted describing paintings' color, texture and other characteristic. Two classification approaches aiming to school and artist classification are implemented with supervised classifiers based on various algorithm, i.e. Bayes Classifier, FLD Classifier and SVMs Classifier. Finally, the performance of individual classifier is assessed using different measurements.