Skip to Main Content
Binary arithmetic codes (AC) with forbidden symbols (FS) can be modeled as finite state machines and treated as variable length trellis codes. In this paper, the relationship between the coding parameters and the number of states is discussed, and a novel adaptive source and channel coding system is proposed based on the relationship. The proposed coding system can adjust adaptively the coding parameters according to the source and channel information so as to achieve the optimal performance. This coding system can be used to transmit images over wireless channel, and the coded bit stream can be decoded by using general list Viterbi decoding algorithm. Experimental results show that this system has some significant performance improvements over previous systems such as the symbol error rate and the computational complexity.