An Algorithm for Optimal Unequal Error Protection Rate Allocation Exploring Granular Channel Rates

In this letter an algorithm is proposed for optimal rate allocation constructed over a wide set of granular channel rates obtained from rateless codes to provide an ideal adjustment for an unequal error protection scheme. This algorithm is based on dynamic programming for resource allocations with capacity restrictions solved as a knapsack problem; formulated to find among each rate arrangement for a progressive bit stream the one that minimizes distortion and maximizes throughput, that represents the adaptation parameters to optimize the rate-distortion performance by balancing resources. Computational results have shown gains in distortion [peak signal-to-noise ratio (PSNR)] of approximately 20 dB over the maximum admissible SNR variation in comparison to the equivalent equal error protection rate. Enough protection levels have been defined to establish the best capacity convergence and algorithm efficiency.