本文提出了一种基于H.264/AVC压缩域的高效全局运动估计算法.由于H.264采用了多种新的视频压缩编码技术,使得其压缩码流的运动矢量(MV)场中包含大量噪声运动矢量,可参与全局运动估计的运动矢量相对较少.噪声运动矢量这里指的是与全局运动不相符的运动矢量.为了降低噪声运动矢量的影响、提高全局运动估计的精度和效率,在全局运动估计过程中就必须对噪声运动矢量进行有效的去除.本文利用H.264码流中的运动矢量预测残差(MVD)和运动矢量预测参考块的信息,对噪声运动矢量做了有效的去除.实验的结果验证了本文提出的全局运动估计算法的计算效率和精度.
An efficient global motion estimation algorithm on H.264/AVC compression domain is proposed.Due to many technologies of video compression coding used in H.264/AVC,on H.264/AVC compression domain,the motion vector field contains a lot of noise blocks for global motion estimation.The valid motion vector number for global motion estimation is limited.In order to reduce the influence of noise motion vectors and improve accuracy and computational efficiency of global motion estimation,it is necessary to eliminate noise motion vectors during global motion estimation.In this paper,the noise blocks were eliminated efficaciously by using information of MVD(the difference between the current vector and the predicted vector) and reference block for MV prediction in H.264 bits stream.The extensive experiments show that the proposed method is efficient and robust in terms of both computational complexity and accuracy.