Abstract In recent years, information technology and computer internet is developing fast, there are many opportunities for us to encounter the image information in our lives。 We have to use those pictures sometimes, so that people need an efficient way to get the pictures which they want。
Therefore, how to get the picture quickly and accurately aroused people's interest 。 Based on this concept, people have done continuous exploration, they pay attention to the several aspects of image features。 For example, they look for image depend on the color of pictures。 And for texture characteristics, they look for pictures depend on the fringe 。 All of this form the Content-based image retrieval technology。 Between all of the characteristics, the color is the most direct characteristics to arose people's attention 。 It's application is being used widely。
This paper focuses on the color characteristics, and make an effecting retrieval system based on color feature。 Choose the color histogram-based approach as basic method。 It is important to select the color space。 Because HSV color model comform to the human eye perceives, so we choose it。 We use non-intervals method to form a feature vector in the quantization part。 At last similarity measure function is being used to retrieve the target image。
Key words: Image retrieval, Color features, Color histogram, Similarity distance