On modeling and turbo compression of grayscale images in the Slepian-Wolf context

In this paper, we frame the grayscale image compression problem according to the Wyner-Ziv problem. We divide the compression process into two separate stages, specifically scalar quantization and Slepian-Wolf coding of correlated discrete random signals. We first explore models of grayscale images leading to image compression scheme based on the Wyner-Ziv problem. Consequently, we develop such a scheme using optimal non-uniform scalar quantization and serial turbo-coding with Latin Square based encoders and report on the obtained simulation results.