一种二进制算术码的软解码算法

算术码是一种高效率的熵编码方法,广泛应用在多媒体压缩中.但是,算术码的高压缩性能也导致其对信道传输中残存的误码极其敏感,存在误码扩散现象.本文将算术码解码过程表示为一个有限状态机(FSM),利用最大后验概率估计准则(MAP)进行序列解码,FSM提供先验信息估计和差错检测功能.通过对独立信源和JPEG2000编码图像的实验表明,该解码算法降低了二进制算术码的序列差错率,提高了解压缩图像的质量.

A Soft Decoding Algorithm for Binary Arithmetic Codes

Arithmetic coding is widely adopted in the state-of-art multimedia compression schemes due to its high coding efficiency.Unfortunately,arithmetic codes are extremely vulnerable to channel errors because of the error propagation property.In this paper,arithmetic decoder is modeled as a finite state machine(FSM),and a sequential decoding algorithm is contrived by utilizing maximum a-posteriori(MAP) estimation rule.The FSM is responsible for a-priori probability estimation and error detection.Simulations with memoryless sources and JPEG2000 images show that the algorithm improves the system performance with reduced sequence error rate and enhanced quality of reconstructed images.