Image Retrieval Based On Color And Spatial Characteristics Of The Study

Color character is always one important vision content in the content based image retrieval. The expression of image color, the extraction of the color character and the mathing algorithms of image likeness based on color character was analyzed and discussed in this paper.Fist, two color spaces(RGB and HSV), three matching algorithms of color histogram(histogram intersection, the distance of OU SHI and proportion distance) and four color histograms methods(tradition histogram, acumulation histogram, local acumulation histogram and quantifying color insector) were dicussed in this paper, the retrieval effects were contrasted through the experiment.Second, aiming at the characteristics of color histogram which were considering the color distributing in the whole image merely but not considering the color distributing in the local area of the image, considering the comparability between two same colors merely but not considering the comparability between two different colors and considering the distributing of color set in the image merely but not considering the shape, size and position of the image object, two improved methods weer brought forward in this paper: one is image retrieval based on Color-Space 2D Histogram, the other is the image retrieval based on Dominant Color and Dominant Color Moment. The former is mainly considering the displacement information of color distributing in the image, the latter is mainly considering the statistic information of color distributing in the image.In the end, three integration methods which are combining Dominant Color and Color-Space 2D Histogram were brought forward in this papar, the result of the experiments indicated that the integration methods could obtain good effect.