Adaptive polynomial interpolation for noise replacement

Abstract

In this paper, a noise replacement scheme based on adaptive polynomial interpolation (NRAPI) is proposed, where the perfect noise detection is assumed. In the NRAPI, the zero-order and the first-order polynomial interpolation schemes are adaptively applied in the noise replacement according to the noise density. It is observed that the proposed zero-order noise replacement scheme generally has better contrast and visual quality when the noise density is high and that the proposed first-order noise replacement scheme has better visual quality for cases of low and mild noise density. Consequently, the proposed NRAPI combines the zero-order and the first-order noise replacement schemes for better performance. The proposed NRAPI is justified by several images and compared with two newly reported noise replacement schemes. The simulation results indicate that the proposed NRAPI has better visual quality of restored images than those from the zero-order and the first-order noise replacement schemes, as expected.