Multi-level 3-D rotational invariant classification

A two-level 3-D rotational invariant classification is developed based on the fractional differencing model. In the first level, classification is done with a fractal scale, and in the second level, textures are classified in detail with additional frequency parameters. Because of the properties of the fractal scale and multi-level procedure, the proposed 3-D rotational invariant classification scheme reduces the processing time and gives enough accuracy to the classification simultaneously. As a result of a series of experiments involving the differently oriented samples of natural textures, it is concluded that these combined features allow this multi-level classification method to have a strong class separability power for arbitrarily oriented 3-D texture patterns.