A Method of Image Retrieval with Color Names
一种利用颜色词检索图像的方法

The contents of image include low level visual features and high level semantic in image retrieval based on content. Generally, the former is data obtained from image analysis and the latter is describing concepts in nature language, but they are not related to each other simply. The paper presents a trial with combining the main colors of color image and color names in Chinese language in color image retrieval. A new color model and a method of calculating the similarity between two colors are proposed after a few of color models are analyzed, on which a method of extracting the main colors of an image is given. The color names are discussed and a method of retrieving color images with color names is proposed. The preliminary retrieving results obtained on trial are discussed, which shows that the method is quite helpful in color images retrieval.