Design And Realization Of Computer Aided Ink Painting Classification System

These years, with the rising prosperity of China art market, ink paintings, as a very China characteristics art, was extremely sought after by everyone. At the same time, the market appeared a large number of ink painting fakes; they not only disrupt the art market and will greatly interference the learning and art heritage for Chinese ink painting. Therefore, this topic in-depth study the basic characteristics and the related algorithm of the ink painting, we develop a computer aided Chinese ink painting classification system to classify genuine and fake Chinese ink painting.In ink mark recognition of Chinese ink painting, this system uses the calculation methods of color histogram and color moment to identify markings and color mixing; in drawing recognition, use of strokes and stroke the disturbance degree elevation with the center of gravity to solve the word problem in the style of quantitative recognition for withdrawals in words; in identifying the seal, use of Lab color difference formula to solve the interference problem of ink color and background color, at the same time, using the OPTA algorithm to extract the seal skeleton. In classification algorithm, this system using the K-nearest neighbor (K-Nearest Neighbors, referred to as KNN) algorithm, after in-depth research and repeated sampling, found that the elements of ink painting has the very big difference on the contribution of the classification system. In order to solve the problem of the classification error problem caused by ink painting its own characteristics, we propose a weighted KNN algorithm, through the weighted arithmetic of each element, to improve the accuracy of the classification system.