Analog Error-Control Coding Based on Dimension-Expanding Shannon Mapping for Robust Image Communication

Analog transmission has the advantage of allowing the quality of recovered data to vary gracefully as the channel conditions vary. For the purpose of robust image communication, two analog error-control coding schemes based on dimension-expanding Shannon mapping are discussed in this paper, including the encoding processes, decoding processes and joint source channel optimizing design methods. In addition, a method of getting the upper and lower bounds of the theoretically optimal coder is presented. In comparison with the optimal coder, these two analog error-control coders are easy to implement and simulations show that they both perform well and are robust to channel and source mismatch. It can be concluded that they are both good candidates for image subband coding to protect important source messages and suitable for robust image communication where channel conditions vary.