The Image Retrieval Based on the Hybrid Algorithm of the Primary Color and Color Layout Descriptor

The image retrieval technology based on content is become a hot research field with the images used widely. The primary color and color layout are both the characteristic in the international standard.Each characteristic factor has its advantages and disadvantages. In this paper, image retrieval based on the hybrid algorithm is proposed in order to avoid their shortccomings. The experimental results show that the better retrieval effect is obtained by using the hybrid algorithm. INTRODUCTION Although there are a lot of colors in an image, the experimental results show that most of images can be used by several or a dozen colors to describe the local color features. The primary color descriptor based on the above ideas is proposed. A set of representative color is specified in the area of any shape, and it is the application based on color. The area can be a whole image or an arbitrary area. The primary color descriptor and the color layout of the standard are utilized comprehensively in this paper. The image retrieval system contained 300 sample is verified. The results show that good retrieval effect is obtained by using the color and the texture feature. PRIMARY COLOR DESCRIPTOR HSV model can be better reflected the perception and the identification ability to color. The H is tone, S saturation, V brightness in the model. The change of each color component can be sensed by HSV model, therefore image similarity comparison based on color is suitable very much. The image in RGB space can be conversed easily to HSV space by a certain formula. The color of HSV space is quantified usually in order to reduce the amount of calculation. The 166 d or 72 d quantization is used in accordance with human visual perception. The formula is such as (1) below:      > < ≤ = otherwise v s h f v s v s h g v q ), , , ( 1 . 0 & 1 . 0 ), , , ( 1 . 0 , 0 (1) Here, ) , , ( v s h g is such as (2) below: