摘要
本文首先提出了一种新的运动估计算法,它利用当前帧中已经译码的相邻图像实现当前图像块的运动估计,避免了运动矢量的编码和传输;然后提出了一种符合图像编码要求的图像分割算法,该算法能够较好地从图像中提取影响图像质量的关键区域.在这两种新算法基础上,本文最终实现了一种性能较好的视频编码算法.
This paper proposes a new motion estimation algorithm, which uses the neighboring decoded image of the current frame and the previous reconstructed frame to estimate the motion vector of the current image block. The encoding and transmitting of motion vectors are not needed because they can be recovered in the receiver. The paper also presents an image segmentation method suitable for video coding, which can extract the important image regions, such as the moving objects.The above two algorithms are used to implement a video coder with good performance.
出处
《计算机学报》
EI
CSCD
北大核心
1997年第12期1081-1087,共7页
Chinese Journal of Computers
基金
国家863项目基金
自然科学基金
博士后基金