Efficient adaptive intra-field deinterlacing algorithm using bilateral filter

Abstract

In this paper, we introduce an efficient adaptive intra-field deinterlacing framework using bilateral filter with consideration of pixel similarity between the interpolated pixel and the neighbor pixels. In the proposed framework, the low resolution image is first preprocessed using the simplest intra-field deinterlacing method, line average, due to its high efficiency and good performance, and then is decomposed using the fast bilateral filter into the detail and smooth layer which represent the small and large scale features, respectively. We process smooth layer and detail layer separately. For detail layer, instead of estimating the edge orientations with some candidate directions as previous intra-field deinterlacing methods, such as edge-based line average, efficient edge-based line average etc., we propose an adaptive but efficient deinterlacing method with consideration of pixel similarity, which avoids considering directional difference measures using limited candidate directions, resulting high efficiency. Compared with existing deinterlacing algorithms, the proposed algorithm improves peak signal-to-noise-ratio while maintaining high efficiency.