mohammad shams esfand abadi - department of electrical engineering shahid rajaee university
ali mahlooji far - department of electrical engineering tarbiat modares university

in this paper a new two-dimensional 2-D adaptive filter algorithm is proposed . The algorithm is based on the matching pursuit technique and is characterized by optimally update only one filter coefficient in every iteration and optimally determining which coefficient to update . an efficient implementation of the algorithm is presented and its computational complexity is treated in detail. the proposed algorithm is illustrated and compared with the 2-D least mean squares LMS algorithm for the applications of image restoration and two -dimensional system identification.

adaptive filtering, least mean square, matching pursuit, image restoration, system identification