Window Averaging Method to Create a Feature Victor for RGB Color Image?

Various applications now are dealing with huge data bases containing various images with different kinds and their own semantics. Color images usually have a huge size, thus they need a lot of time for processing tasks such as image retrieval, image transmission and image recognition. In order to reduce color image processing time and memory space size needed to accomplish the processing cycle an efficient method of features extraction will be proposed. This method will allow the user to create a feature victor for each color image; this victor will be used as key for color image retrieval or recognition. This key will optimize the processing task because it will have a very small size comparing with the color image size. The proposed method will be tested, implemented and the results of implementation will compared with other method in order to show the efficiency of the proposed method.