The Personification Research Of The Peking Opera Painted Faces

Peking Opera Painted Faces(POPF)is a colorfully traditional artistic expression form.It can convey the personality traits of appropriate Peking Opera role to the audience effectively.So far,the researches of the personality traits of appropriate Peking Opera roles are mainly confined to the field of art,and there is ambiguity and uncertainty in the description of personality traits conveyed by the POPF,because it relies on the experience and perception of the judge.In this paper,we proposes a method combining with the big five model of personality to analyze the personality traits conveyed by the POPF quantitatively.First of all,we collected 300 subjective ratings of POPF's personality traits based on the questionnaire forms from college students.And the questionnaire forms were designed with the big five personality scale bipolar emotional adjectives.Then we established the 5-dim personality factor space by analyzing the factors of subjective ratings data.Secondly,the SVM algorithm was used respectively to construct the 5-dim personality factor recognition models according to two kinds of the POPF image features(low-level image features and the features learnt by the convolution neural network pre-trained by sparse coding).Comparing these two kinds of the personality factor recognition models,we selected the better one to construct the Peking Opera personification retrieval system.In this paper,the quantitative description of POPF's personality traits is preliminary realized,which is helpful to the inheritance and development of the Peking Opera culture.The main contributions of this paper are as follows:(1)Constructing a library of POPF's personality traits.We constructed the POPF library by accessing to information online and offline,which includes 700 POPF images,and collected 300 effective subjective ratings of POPF's personality traits from college students based on the questionnaire forms designed with the big five personality model.Then we constructed a library of POPF's personality traits.(2)Using SVM algorithm and POPF image features to construct the POPF's personality factors recognition models.POPF image features include low-level image features and the features learnt by the Convolution neural network pre-trained by sparse coding.Put the semantic description of POPF's color and the edge direction histogram feature of POPF as the low-level image features.Finally,the SVM algorithm was used to model the POPF's personality factors recognition modes.(3)Constructing the POPF personification retrieval system.We design the Peking Opera personification retrieval system based on the POPF's personality factors recognition models.The system implemented the retrieval among POPF roles,POPF images and POPF's personality.The system could be used as a new method in the propaganda and education of POPF culture,and it will be helpful to the inheritance and development of the Peking Opera culture.