Structural edge detection based on fractal analysis for image compression

Addresses the problem of structural edge detection and discusses the possibilities of application of this edge information to image compression. For the purpose of detecting an edge efficiently, the authors first divide the image into appropriate nonoverlapping blocks and estimate the fractal dimension of each block by using the blanket-covering method. Secondly, the edge points are efficiently detected by using a wavelet transform where the dilation parameter is controlled based on a fractal dimension. The proposed method was tested on a real image using two wavelet functions, and the results are shown.<>