Enhancement of interlaced images by fuzzy reasoning approach

Abstract

This paper investigates a new interpolation technique based on content adaptive interpolation with fuzzy reasoning. The proposed technique reduces image artifacts in regions where moving object cannot be well traced by motion estimation. Our proposed method consists of two parts: content classification and fuzzy weight assignment process. The spatial deinterlacing methods have been studied excessively in literature. The edge based line average method yields relatively good performance with low complexity. On the basis of the above method, six useful measurements are obtained within the 5-by-3 spatial window to prevent false edge direction decisions in computing the direction where the interpolation is to be operated. The simulation results indicate that the developed method can generate high quality intra-interpolation picture for interlaced images.