Web Image search re-ranking, is an effective way to improve the results of web image search, which has been adopted by commercial search engines like Bing and Google, etc. When we give a query keyword, a pool of images are retrieved first based on query information. We ask the user to select a
image from the pool of images, so that the remaining images are re-ranked based on their visual similarities with the selected image. A major problem is that the visual similarity features do not well coordinate with images’ semantic meanings. We need to match images in a semantic space which uses reference classes closely related to the semantic meanings of images. In order to measure similarity FCTH (Fuzzy Color and Texture Histogram) and CEDD (Color and Edge Directivity Descriptor) are used. The proposed features are appropriate for accurately retrieving even in distortion cases such as deformation, noise and smoothing. It is tested on a large number of images selected from proprietary image databases or randomly retrieved from popular search engines. Low computational power is needed for its extraction. Image understanding is widely used in many areas like satellite imaging, robotic technologies, sensory networks, medical and biomedical imaging, intelligent transportation systems, etc. But it is difficult by traditional image processing.