Motion-adaptive Scan And Its Application To Video Compression Based On Multi-dimensional Vector Matrix

With the development of multimedia technology and communications technology,a variety of applications of image and video has become more and more frequently inpeople’s daily life. Thus, the explosive increasing data of image and video hasproposed high requirements not only on existing transmission bandwidth and storagespace, and also on the technology of image and video compression. Under this kind ofexternal power, it has formed many famous compression coding international standards.And these standards greatly promote the progress of multimedia informationtechnology. Whereas they have to face the challenges of further improving thecompression ratio and reconstructed video quality.The goal of compression coding technology pursuits is to describe videoinformation with the bit rate as small as possible under the guarantee a certain qualityconditions of the reconstructed video. Based on Shannon Information Theory, thetraditional compression technology regards image and video data as a source describedby statistical probability model, which consist the amount of information (sourceentropy) and information redundancy. This compression coding technology based onwaveform is a great success, such as MPEG-1，MPEG-2,MPEG-4，H.26X and so on,which formed a tremendous influence to multimedia information industry. Thesecompression technology used motion compensation prediction and discrete cosinetransform hybrid coding (DCT/MCP (Discrete Cosine Transform/MotionCompensation Prediction)) as main feature. In general, accord the correlation of theinter-frame predictive coding mode to judge. Using frame differential predictivecoding remove spatial redundancy or using inter-frame motion compensationpredictive coding remove temporal redundancy. Then transform the predictive error(usually using DCT). However, this judgment resulted that inter-frame and intra-frameredundancy can not be solved on the same time. If the encoded video-object beyond three-dimensional, traditional compression coding can not eliminate the redundancy ofthe various dimensional completely. Springing up of multi-dimensional compactiontechnology reasoned to solve this subject.Multi-dimensional vector matrix proposed by Sang Aijun supply mathematicalguidelines for research to multi-dimensional data compression technology integratedtemporary redundancy and spatial redundancy to unified mathematical model take intoeliminate, that change the confusing situation of multi-dimensional operational rules,although this architecture still in the initial establishment.According to compression of temporary redundancy from video data, we study onmulti-dimensional vector matrix compaction technology and global constanttranslation motion in MD-DCT, method have been proposed to estimate point oflocation area in the multi-dimensional space coordinate system grouped around energyof transform coefficients, then according to this point established scheme scan oftransform coefficients named Motion-adaptive Scan.The effectiveness and progressiveness of the new reproach of scan method in thispaper is verified via simulation experiments, conceived the future research orientationrelated technology.