Research And Implementation Of Color Naming Method Under Different Illuminations

The name of color is the language label by which human can contact with colors.Machine simulates the color perception system of human to identify colors, and label the color's name.That can help to identify the color of specific object.This kind of method has many practical applications,such as color naming methods combine with other content-based search methods to make results of search more specific,more accurate and more close to the way of human beings' judgment.Color naming algorithm is the basis of the color recognition;Different lighting conditions and the choice of learning data sets directly affect the accuracy of latter processes,such as the color tagging and the processes which based on the result of color naming.According to previous studies,some human factors and external factors may bring in a certain degree of error.In order to avoid or reduce the impact of these factors, the paper offers an automated method of color naming - Color naming method under the different illuminations.Firstly,we do the researches of the basic theory and some classic algorithms of color naming;show the advantages and disadvantages of each algorithm by comparison.Secondly,we offer color naming method under the different illuminations,using color constancy algorithm make color training set contain more relevant color information.Image segmentation technology which has been brought into the color naming method focused on making the color information of learning more concentrated, making learning process more effective.Through experiments we verify the feasibility and efficiency of such color naming method.Finally,the details of algorithm implementation and the flow chart of the algorithm are displayed in this paper for the experiment of different illumination conditions.We do an objective analysis of the data and results of experiment,compare and conclude with existing algorithms,then show the images of color naming according to the comparison.Experiments shows that the proposed algorithm by this paper can adapt to different light requirements of the real image set,color constancy algorithm of color naming process improve the adaptation of color naming among different illuminations.Through the relevant technology of image segmentation algorithm,we can name color with lower error rate,and get better result of color naming.