Rotation invariant texture analysis using radon and fractional fourier transform

In this paper we propose a new rotation invariant feature descriptor for texture classification and clustering via first identifying the so-called principal direction of textures with the well-known Radon transform and then extracting features defined by the fractional Fourier transform of different order from the rotated textures along their principal direction. The performance of the proposed method is evaluated using different kind of texture sets. Results show the advantage of the proposed method over some existing algorithms.