A new two-dimensional fast adaptive filter based on the Chandrasekhar algorithm

In this paper, we present a new fast algorithm for two-dimensional (2-D) linear adaptive filtering using the fast Chandrasekhar equations. Using the analogy between the multichannel linear model and the 2-D one, we transform an image into multichannel sequence and we extend the fast Chandrasekhar adaptive multichannel filtering algorithm to the 2-D case i.e. image filtering. The performance of the new 2-D adaptive filter is tested by using this filter to estimate the coefficients of a 2-D Moving Average (2-D MA) model of an unknown system. Furthermore, an application on adaptive noise cancellation of images is proposed throw a 2-D adaptive noise canceller based on the 2-D Chandrasekhar fast algorithm. Simulation results prove the superiority of the new 2-D Chandrasekhar filter comparing to similar approaches for image model identification.