Balanced Multiple Description Coding for 3D DCT Video
Supriya Sawant, Donald Adjeroh · IEEE Transactions on Broadcasting · 2011
We propose a balanced multiple description coding framework using rate-distortion splitting for three dimensional DCT (3D DCT) coded video data. The proposed algorithm produces correlated and balanced descriptions, with balancing in terms of bit rate and receiver distortion. Our description splitting and balancing algorithm takes advantage of the structural properties of the 3D DCT video volume and the method of shifted compliment hyperboloid used in scanning and quantization of 3D DCT coefficients. Given the available bit rate and redundancy our simulations show high PSNR values for the reconstructed video. The proposed video codec requires a simple implementation, has a low complexity, and produces a high reconstruction quality. Comparative results with other schemes are included.