Optimization of Hierarchical 3DRS Motion Estimators for Picture Rate Conversion

Abstract

There is a continuous pressure to lower the implementation complexity and improve the quality of motion-compensated picture rate conversion methods. Since the concept of hierarchy can be advantageously applied to many motion estimation methods, we have extended and improved the current state-of-the-art motion estimation method in this field, 3-D Recursive Search (3DRS), with this concept. We have explored the extensive parameter space and present an analysis of the importance and influence of the various parameters for the application of picture rate conversion. Since well-performing motion estimation methods for picture rate conversion show a tradeoff between prediction accuracy and spatial motion field consistency, determining the optimal tradeoff is an important part of the analysis. We found that the proposed motion estimators are superior to multiple existing techniques as well as standard 3DRS with regard to performance at a low computational complexity.