The Research On Internet Transmission-Oriented Error Resilient Image Coding

With the expeditious development of Image Communication Applications, robust transmission of high compressed images over unreliable Internet has become one focus in image compression coding.Given for ideal channel, existing image coding algorithms and frameworks are concentered on data organization and quantization to pursue high compression efficiency. The ignorance of actual Internet characteristics results that they are short of transmission robustness and cannot ensure both high compression efficiency and high reconstruction quality. To get a balance between the paradoxical aspects, this paper researches on efficient error resilient image coding over low bandwidth and packet loss-prone channel transmission.In order to guarantee high compression efficiency and high reconstruction quality, character of packet-loss in Internet and lack of error-resilient for previous algorithms are firstly studied. Based on the analysis, this paper sums up applied principles of robust image communication over error-prone heterogeneous Internet, in which the concept of globality and Independency are emphasized.Aiming at packet loss in Internet and following the above principles, this paperproposes a whole solution-efficient and robust Joint Sink-Source image codingframework(J2SC), in which both compression efficiency and reconstruction quality are considered. J2SC combines error resilient algorithm in source with error concealment technology in sink to improve image loss-oriented transmission robustness. A global pseudo-random sampling scheme based on statistic distribution of wavelet coefficients is included to overcome error sensitive caused by dependent coding. Also, this paper builds a duplex pyramid model referred to Laplacian Pyramid, which makes image reconstruction quality better.There are detailed comparisons with existing image coding algorithms and JPEG2000 in the end. The MATLAB simulations conclude that J2SC not only can meet the requirements of high efficiency and flexible scalability, but also can provide excellent packet-loss resilient performance adapting to the applications of image transmission over Internet.