Intra-field selective deinterlacing algorithm considering the fine edge direction

Abstract

This paper proposes a selective intra-field deinterlacing algorithm scheme by using a detailed edge direction classification method. It is base on a mixture of different well known approaches: i) M-ELA method to predict the edge and interpolate the missing pixels, ii) Sobel mask for obtaining more accurate edge detection, iii) using the ratio of vertical edge and horizontal edge, the approach selects the most appropriate method among LA method, M-ELA method and doi: method. The experimental results show that this proposed algorithm has outperformed the existing deinterlacing algorithm in terms of both objective and subjective evaluation.