关键词: DCT域 下采样 视频解码 MPEG

Abstract:

To improve the video decoding ability on the multimedia processing capability limited terminals, an effective algorithm is proposed to implement the DCT-domain downsampling video decoding. The algorithm performs downsampling inverse DCT by taking advantage of the symmetric convolution and centered-ene  rgy properties of DCT, and exploits the downsampling motion compensation with proper motion vector approximation, interpolation and reconstruction. The  computation complexity analysis shows that the downsampling inverse DCT, the pixel interpolation and the image reconstruction saves about 50%, 50% and 7  5% computing operations respectively compared to the regular ways, thus improves the overall decoding efficiency by about 60%. The algorithm has been implemented and tested on the XSeale embedded platform and has achieved good performance.

Key words: DCT-domain, downsampling;video decoding, MPEG