Scalable multiple description coding with side information using motion interpolation

Problems with modern video delivery systems is that decoder fails to decode video when it is transmitted over error prone channels, due to varying characteristics of a wireless channel or congestion in wired networks. Thus introducing mismatch into the decoded video stream. Scalable video coding (SVC) encounters such a problem as the loss of a lower layer prevents all higher layers being decoded, even if the higher layers are correctly received. Multiple description coding (MDC) can help solve such a problem. We propose a scalable MDC architecture that creates an even and odd streams with residual information regarding the other stream. Using such a method creates redundancies that can improve the decoded reconstruction of a frame however increasing the bit rate of a given stream. To solve this a motion interpolation technique is used to improve the construction of side information thus reducing the bit rate but improving the frame quality. Experimental results show that motion interpolation scheme can improve the frame quality by 6 dB over a pixel based interpolation scheme for a number of video sequences Experimental results also show that this method can give 40% saving in bit rate when compared to a pixel based interpolation. Using such a method makes the system a viable target for mobile communications.