A New Image Retrieval System Based on CBIR

content based image retrieval from large databases has turn into a field of broad attention these days in much application. In this paper we present new image retrieval system that employs color as image features to illustrate the content of an image region. Our work presents an evaluation of the proposed CBIR systems: CM, DM, HA, RGBC and ED. These methods boost the retrieval precision of our system; we merge the features extracted from image regions, with features extracted from the whole image, which are color histograms. Our proposed system has the advantage of increasing the retrieval accuracy. The testing assessment of the proposed system is founded on a WANG’s color image database. From the experimental results, it is marked that proposed system achieves appreciably better compared with other existing systems. Keywords—Content Based Image Retrieval (CBIR), histogram-based image retrieval, color quantization, color model, precision, Histogram similarity measures.