摘要
提出了一种帧存储器容量受限条件下编码视频码率变换结构 .在这种结构中 ,将压缩的参考图像存入帧存储器中以节省存储空间 ,同时通过截除适量与高频 DCT系数对应的码字以适应任意容量的帧存储器 ,而运动补偿直接在 DCT域实现 ,避免频域与空间域转换所需的运算量 .这种结构的性能随配置帧存储器容量的增加呈单调递增的关系 。
A memory scalable architecture for rate scaling of coded digital video is proposed.In this architecture,the reference anchor frame is compressed and scaled to fit the arbitrary size frame store.And the motion compensation is implemented directly in DCT domain to avoid the data conversion between the frequency domain and the spatial domain.The advantage of this architecture is the memory saving and the memory scalable,and the quality of the scaled bitstream is propotional to the size of the allocated memory size.It provides an approach to generate a converted bitstream,which can offer satisfactory picture quality when the size of available memory is insufficient.
出处
《天津大学学报》
EI
CAS
CSCD
2000年第2期162-166,共5页
Journal of Tianjin University(Science and Technology)
关键词
编码视频
码率变换
帧存储器容量
多媒体通信
rate scaling
Lagrangian optimization
motion compensation in DCT domain