Dynamic rate allocation with variable quantization in multi-sensor Wyner-Ziv video coding systems

In this paper, we present a novel algorithm for dynamic quantization in distributed Wyner-Ziv video coding. In contrast with previous work where the quantization parameter is fixed and a feedback channel is used, our proposed technique relies on theoretical calculations to jointly determine the number of quantization levels along with a suitable compression rate for each video frame. It employs a cross-layer approach that dynamically allocates unequal transmission rates for different users by taking into account the amount of motion in the captured video scenes on one hand and the transmission conditions for each sensor on the other. The application of this algorithm in a wireless video sensor network shows a significant improvement in the system performance when compared to a traditional system that allocates equal channel resources with a fixed quantization parameter.