A new image comparing technique for content-based image retrieval

In this paper, we propose an image comparing technique that can be used for sorting images in large image databases for the sake of obtaining fast retrieval performance. During the indexing phase, a distance value is extracted for the image to be indexed that represents the distance between this image and a reference image in the database. The extracted value is used to insert the newly indexed image in its proper position in the database so that the entire list is always sorted. In the search process, we use the distance value to narrow down the scope ofsearch and then use another comparing technique to obtain the Jinal results. Experiments showed that the proposed approach performs faster than previously reported techniques and eliminates false positive results as well..