Translation Scaling and rotation invariants of 3D Krawtchouk moments

The discrete orthogonal moments such as 3D Krawtchouk moments are effective for 3D object recognition. The translation scaling and rotation invariants of 3D Krawtchouk moments are attained either by normalizing the 3D object or with using a combination of the corresponding invariants of 3D geometric moments. However, the extraction of these moments is not founded on these polynomial. In this work, we suggest a new approach to extract the translation scaling and rotation invariants of 3D Krawtchouk moments directly from these polynomials. The performance of the suggested approach is tested using 3D object. Numerical simulations show that the 3D Krawtchouk moments are invariant under 3D translation scaling and rotation. Moreover, the rapidity of the suggested approach is faster than traditional methods.