Motion estimation searching method for cross hexagon

Abstract

The invention relates to a novel cross hexagon search method, a small cross mode is firstly used for carrying out the pre-search to find a minimum block matching distortion (MBD) point; a large cross search mode is constructed by taking the MBD point as the center, and the MBD point is found; then the hexagon search is started in the following manner by taking the MBD point of the large cross mode as the center: a large hexagon is firstly searched, if the MBD points is in the center, the search is carried out by using a small hexagon mode, the found MBD point is the final minimum block matching distortion point, otherwise, the large hexagon search is continued. The cross hexagon search adopts a halfway stopping technology and significantly improves the search speed of static and semi-static blocks. The improved partial distortion principle greatly reduces the computational complexity under the situation of not affecting the distortion degree. The experimental results show that compared with the hexagon search and the cross search, the cross hexagon search can reduce the search points by 45 percent and 28 percent respectively under the situation of reducing the signal noise ratio a little or even not reducing the signal noise ratio, and compared with other popular block matching motion estimation methods, the method has faster search speed and smaller distortion degree.

Description

technical field [0001] The invention relates to a processing method in video compression, in particular to a search method in fast block matching motion estimation in video compression. Background technique [0002] Because video sequence images have strong correlation on the time axis, motion estimation (ME) and motion compensation (MC) techniques can effectively reduce temporal correlation, so this technology is widely used in various video compression coding schemes. Motion estimation is used to estimate the displacement of the object to obtain the motion vector; motion compensation adjusts the displacement due to motion in the previous frame according to the obtained motion vector, so as to obtain the predicted frame as close as possible to the current frame. It can be seen that the more perfect the motion estimation algorithm is, the more accurate the estimated motion vector is, and the better the performance of motion compensation will be, so that the prediction error ...

AI Technical Summary

Problems solved by technology

Among all the search algorithms, although the full search algorithm has the highest accuracy, its huge computational complexity makes it unsuitable for real-time applications.

The three-step method reduces computational complexity by limiting the number of search positions, which is not conducive to estimating small motions and is easy to fall into a local minimum

Images