Cascade fuzzy filters for impulse and random noise cancellation

Abstract

This paper proposes the design of a multipass fuzzy filter for noise cancellation, and considers its application to video deinterlacing. The proposed fuzzy filter is developed to reduce noise with random and impulse characteristics, and to be used in real-time deinterlacing. We first present an impulse noise cancellation filter followed by random noise removal filter. The proposed technique can integrate conventional deinterlacing methods with the proposed denoising operation. The algorithm is tested on video sequences with impulse and random noises. The simulation results prove that the proposed method suppresses noise efficiently. In addition, we trace optimal parameter values for fuzzy membership function.