An efficient chaotic source coding scheme operating on variable-length blocks is proposed.With the source message represented by a trajectory in the state space of a chaotic system,data compression is achieved when the dynamical system is adapted to the probability distribution of the source symbols.For infinite-precision computation,the theoretical compression performance of this chaotic coding approach attains that of optimal entropy coding.In finite-precision implementation,it can be realized by encoding...","An efficient chaotic source coding scheme operating on variable-length blocks is proposed.With the source message represented by a trajectory in the state space of a chaotic system,data compression is achieved when the dynamical system is adapted to the probability distribution of the source symbols.For infinite-precision computation,the theoretical compression performance of this chaotic coding approach attains that of optimal entropy coding.In finite-precision implementation,it can be realized by encoding variable-length blocks using a piecewise linear chaotic map within the precision of register length.In the decoding process,the bit shift in the register can track the synchronization of the initial value and the corresponding block.Therefore,all the variable-length blocks are decoded correctly.Simulation results show that the proposed scheme performs well with high efficiency and minor compression loss when compared with traditional entropy coding.