Median Based Wiener Filtering interpolation method for Progressive display using Local Homogeneous Area Selector

摘要

This paper proposes an efficient intra-field deinterlacing algorithms using local adaptive median-based Wiener filtering interpolation method (LAMWF) using Local Homogeneous Area Selector (LHAS). The LHAS is introduced to determine the local homogeneous area where the center unknown pixels belongs to. After using LHAS, a median-based Wκiener filter is proposed to interpolate the missing unknown pixel. Since LAMWF is only implemented on the optimal neighboring pixels in a homogeneous area, which makes it has good capacity of preserving edge and texture details. Experimental results have shown the efficiency of the proposed method with significant improvement over conventional methods in terms of both objective PSNR performance and subjective image qualities. What is more, it has benefit of lower complexity than most of the existing algorithms.