Image feature extraction and segmentation using fractal dimension

After several existing approaches of fractal dimension estimation for image analysis have been introduced, two new approaches are proposed: wavelet energy fractal dimension and morphological fractal dimension. Then we try to obtain variable fractal dimensions of general images such as Lenna instead of typical texture images as done before. Experimental results demonstrate that using different fractal dimension estimation approaches, different instinctive self-similar properties of one image can be achieved, so that they can be used for image feature extraction and segmentation