Analysis of Krawtchouk Moments Rotation Invariance
Chenhui Yang · Modern Computer · 2009
Moments is a shape descriptor based-on region,it has better performance for description of unconnected-region-image and robust against noise comparing to those methods based on contour such as fourier descriptor and chain-code descriptor.Orthogonal moments can glancing be divided into two categories,continuous orthogonal moments and discrete orthogonal moments.Krawtchouk moments is one of discrete orthogonal moments.By comparison,discrete moments is more suitable for digital image due to its discreteness.As other discrete-moments,Krawtchouk moments is not basic invariant with geometric transformation(rotation,scaling and translation),and Krawtchouk moments is not widely spread used in this case.To solve this problem,gives the suitable rotation invariant Krawtchouk moments,by comparing different methods and experiments.