Match searching method and system based on motion search window

Abstract

The invention discloses a match searching method and system based on a motion search window. The method comprises the following steps: specific to a current macro block of a current frame, establishing a motion search window of a first preset size; calculating the matching cost of each initial iteration candidate point, and taking an initial iteration candidate point with a minimum matching cost as an initial iteration point; and determining a target position of match searching from a motion search frame through the initial iteration point, wherein each initial iteration candidate point comprises the current macro block, a point corresponding to the motion vector predicted value of the current macro block, an encoded macro block adjacent to the current macro block, and a point corresponding to the central position of the motion search window. The system comprises a motion search window establishing module, an initial iteration point calculation module and a target position determination module. Through adoption of the method and system, match searching is not only performed close to the current macro block, but a required bandwidth can be kept within a controllable range.

Description

technical field [0001] The present invention relates to the technical field of video coding, in particular to a matching search method and system based on a motion search window. Background technique [0002] Based on the similarity principle in the time domain, blocks in similar regions of different frames have a higher similarity. Therefore, when the encoder processes a macroblock of the current frame, it accesses a certain area of ​​the reference frame to find the most matching macroblock, and achieves the goal of reducing the bit rate by encoding the difference between the reference frame and the corresponding area of ​​the current frame. Since the location of the most matching area cannot be determined, it is necessary to search for a matching macroblock in a larger estimated neighborhood. We call this neighborhood MSW (Motion Search Window, Motion Search Window). Generally, the size of the reference frame is large, and it is usually stored in the external DRAM (Dynami...

AI Technical Summary

Problems solved by technology

This method has a more flexible search range and is better than the first method for scenes with intense motion, but the disadvantage is that each macroblock corresponds to a different search area, resulting in uncontrollable bandwidth consumption

Images