Fine directional de-interlacing algorithm using modified sobel edge detection

Abstract

In this paper we propose a fine directional de-interlacing algorithm, based on detection of an accurate edge direction using modified Sobel operation. Existing deinterlacing algorithms determine the presence of edge in pixel resolution or half pixel resolution, so that they can be highly sensitive to noise and lead to image degradation. Therefore the proposed method first considers rough edge tendency, and then through the modified Sobel edge detection it is adaptively adjusted to elucidate fine edge direction using its tendency. Finally, proposed fine directional de-interlacing algorithm yields accurate results compared with other edge-directed deinterlacing algorithms. Experimental results show that the proposed algorithm performs well with a variety of still images compared with conventional intra field deinterlacing algorithms in the literature.