Application for deinterlacing method using edge direction classification and fuzzy inference system

Abstract

In this paper, we propose an advanced deinterlacing method which uses filters to estimate the edge direction using luminance information. Subsequently, we are able to obtain the luminance values at for missing pixels. The fuzzy logic concept for image processing is discussed with regard to fuzzy membership function representation and fuzzy inference procedures. The fuzzy if-then rules are employed to conduct the determining edge direction. The use of a different membership function for different direction enables the filter to independently characterize separate influences on pixel variation. Simulation results demonstrate that the proposed method has an enhanced performance, both visually and in terms of the peak signal-to-noise ratio, compared with those of conventional deinterlacing methods.