Fast fractal image retrieval algorithm based on contiguous-matches

The compression transform parameters that keep the image approximate constant are recorded in fractal code, which reflect the cross-scale redundancy features between different regions of a natural image. Therefore the application of the parameters to content based image retrieval will achieve effects. In this paper, a new image fast fractal encoding method based on adaptive quadtree is put forward. In addition, an image retrieval algorithm and the distance formula are also given. Along with the substantial increase in retrieval effect, the image decoding quality is also guaranteed. Results show that this algorithm significantly reduces the encoding time of image retrieval, and the accurate-complete rating of which is obviously superior to that of the color histogram method.