Fast algorithms for motion estimation in video sequences using frequency domain

This book is about a thesis defended in 2007 that is interested in the development of algorithms for the estimation of movement between two successive images for video processing and coding applications. Block matching techniques are generally the most widely used. In this context, the best solution from the quality point of view is represented by an algorithm of exhaustive research. However, this algorithm requires a huge computational complexity. Different suboptimal solutions have been proposed in the literature, but an alternative approach to the problem in the frequency domain is still missing. This thesis proposes fast algorithms to accelerate the process of a comprehensive search using different metrics.The approaches introduced, unlike several proposed solutions, are not based on the spatial domain, rather they use the domain frequency. The proposed methods make it possible to provide computation time deterministic and fast while maintaining a high-quality motion estimation equivalent to an exhaustive search.