Multifractal analysis based on multi-resolution of medical images

This paper deals with the problem of obtaining and analyzing texture features in medical images. For this purpose we make use of a method based on the multifractal concept to compute the characteristic vector of fractal dimension (MFanvg) for CT images of normal liver or lung, CT images of abnormal liver or lung, MRI images of them and so on. It was discovered from the results, whether we implement multifractal analysis based on multi-resolution to CT images or MRI images, that each element of the MFanvg had no difference for normal parenchyma images but not for abnormal images. These differences can be testified by the Variance (Var) of characteristic vector MFanvg, of which the Var value of normal parenchyma image is less than 0.01 but abnormal more than 0.01.