A Replacement Based Video De-Interlacing Technique by Feathering Effect Detection and Artifact Agglomeration Index

摘要

Several video de-interlacing techniques have been already developed and each one presents better performance in certain conditions. Occasionally even the most modern de-interlacing techniques create frames with worse quality than primitive de-interlacing processes. This paper validates that the final image quality can be improved if the frames are processed locally and selectively, combining different types of de-interlacing techniques. The proposed strategy is able to identify and locally correct defects on de-interlaced frames with the support of a feathering effect detector. Starting from a de-interlaced frame produced by “inter-field average” method, the defective areas are identified and fully replaced by pixels generated through “edge-based line average” method (ELA). Test results have proven that the proposed technique is able to produce video frames with higher quality than applying a single de-interlacing technique through getting what is good from intra- and inter- field methods.