Research And Implementation On Classification Algorthm With People Of Ancient Chinese Murals Based On Style Characteristics

Ancient Chinese mural is an important part of Chinese art and culture,with the development of Chine history,Ancient Chinese mural including a record of all times of the folk habits,social status and religious content such as names,have very important research value.With the continuous development and the promotion of computer technology direction of image processing,the mural image is also thinking of the digital age,the mural image digital enables researchers to use the computer to help the mural image analysis and research on the mural image.China's relatively intact with large-scale murals images,researchers for the mural image provides a rich research data and materials.Among them,how to find from this mass of material they are interested in the image is an urgent need to solve problem,and classifying the mural image research can help researchers select interested in content,is a good way to solve this problem.While the existing image classification technology is not very practical for the ancient Chinese murals,so this paper speaks of a mural art style of ancient Chinese characters based classification techniques according to the characteristics of different styles in different periods,mural character classification research?Research goal of this paper is dividing these murals into different categories with art characteristics of different dynasties based on the difference characteristics characters murals,headwear,clothing style between murals.This paper divides the process of murals classified mainly into three parts: localizing the goal from image,namely getting the goal of people from the mural painting;Extract the feature from target image,that is,to extract characters like color,texture and contour features from mural;According to the characteristics of the target and eigenvalues to classify the mural.Main contributions are of this article:(1)Proposing a detection method for the aims of the characters in the ancient Chinese murals for murals characters in target detection.Namely to face detection,then to detect human face up and down can be extended with character of target image.(2)This paper proposes a new classification method for murals.The first to get a large number of murals of the target image characters,extract the image color,texture,contour features,according to the average value of these characteristics,class variance,average variance data classification,etc.In this paper,the classification is not strictly divided into different dynasty murals,but is divided into different dynasties style murals,such as the works of Tang Dynasty or Song Dynasty may have obvious Han Dynasty style,can be classified as the Han Dynasty style murals.