A Motion-Adaptive De-interlacing Method Using an Efficient Spatial and Temporal Interpolation

Abstract

This paper proposes a motion-adaptive deinterlacing algorithm based on an edge-based median filter (EMF) and adaptive minimum pixel difference filter (AMPDF) To compensate for missing-motion error, which is an important factor in motion-adaptive methods, we used an AMPDF, which estimates an accurate value using different thresholds after classifying areas of the input image into 3 classes. To efficiently interpolate moving diagonal edges, we used an EMF, which adopted a 5-point median filter using the edge information. Finally, to increase the performance, we adopted adaptive interpolation after subdividing the input image into moving, background, and boundary regions. Simulation results showed that the proposed method performs better than existing methods.