Survey on block-based motion estimation for frame rate conversion

摘要

The basic idea of motion estimation is to divide each frame of an image sequence into a number of non-overlapped blocks, with the assumption that all pixels in a block have the same amount of displacement, then give each block a specificity reference frame. In the specified search range, a block that is most similar to the current block is considered to be a matching block according to a certain matching rule. There are many methods for searching for matching blocks, such as Full Search, Three Step Search, New Three Step Search, Four Step Search, and Diamond Search. These methods effectively improve the search accuracy and make the motion vectors obtained from motion estimation more accurate. This article mainly introduced the basic principles, implementation techniques, advantages and disadvantages, and the optimal and worst search complexity of various algorithms under different search windows of these algorithms. Finally, these algorithms are used to perform motion estimation, and the same method of motion compensation is performed to experimentally illustrate the applicable scenarios of various algorithms.