Anisotropy analysis of textures using wavelets transform and fractal dimension

In this paper, we propose a new method based on texture analysis of anisotropy. Our proposed method based on a combination of fractal analysis with preprocessing using histogram equalization with discrete wavelets transforms (DWT). First, the texture image enhanced using the histogram equalization. Then, the enhanced image rotated with differences angles from 0° to 360° with a step of 15°. After that, we applied the DWT; we have chosen the Daubechies Wavelets (dbn) for each rotated images. This step followed by the fractal analysis using the differential box counting (DBCM) to the approximate image to estimate de directional fractal dimension (FD). Finally, the degree of anisotropy (DA) calculated as the ratio between the maximum and the minimum value of the FD. The originality of our work reside in the use of the Daubechies Wavelets (dbn) in particular the use of approximate image with the fractal analysis by estimating the directional FD and analysis of the anisotropy. The testing and evaluation of our algorithm are carried out using some textures of the Lille INSERM database U 703 that contains two modalities (MRI and CT-Scan) of bone trabecular texture ROI (Region Of Interest) healthy and pathologic and some Brodatz textures.