Optimal Protection for Progressive Image Transmission over Noisy Channels: A General Approach

In this paper, we studied the problem of transmissions of embedded image bit streams over nite state Markov chain (FSMC) channels. A general approach called error sequence analysis (ESA) is proposed to evaluate and optimize the the performance of the received image quality in various channel conditions (Table1). By using ESA, a general source model for the embedded image bit stream called the av erage of the rst error-free run length (FEFRL) is obtained in close-form. Existing results on BSCs and two state memory (Gilbert-Elliott) channels are shown to bespecial cases of the model. F urthermore, there is no constraint on the structure and the number of states of the FSMC channels in the proposed approach. With this source model, the motion (Doppler) e ects on progressive image transmissions in fading channels are rev ealedfor the rst time. The relationships to other approaches are also clari ed.