CLASSIFICATION OF ROTATED AND SCALED TEXTURED IMAGES USING INVARIANTSBASED ON SPECTRAL MOMENTS

This paper describes a classification method for rotated and scaled te xtured images using invariant parameters based on spectral-moments. Al though it is well known that rotation invariants can be derived from m oments of grey-level images, the use is limited to binary images becau se of its computational unstableness. In order to overcome this drawba ck, we use power spectrum instead of the grey levels to compute moment s and adjust the integral region of moment evaluation to the change of scale. Rotation and scale invariants are obtained as the ratios of th e different rotation invariants on the basis of a spectral-moment prop erty with respect to scale. The effectiveness of the approach is illus trated through experiments on natural textures from the Brodatz album. In addition, the stability of the invariants with respect to the chan ge of scale is discussed theoretically and confirmed experimentally.