Study on the Variable Rate Coding of Moving Image Sequence in the Next Generation Communication Network

异步传输模式（ATM）因其多样性和可用性而成为目前最好的信息传输方式之一。在ATM网络中，通过统计复用可以实现各种类型的通信连接，有效地利用网络资源。在使用ATM的运动图像通信系统中，用户可以不受比特率的限制接入网络。因此，必须从根本上改变编码方案的火焰工作，从而计划从恒定比特率转向恒定图像质量。在本研究中，我们将高效的图像编码问题分解为两个功能块，一个是构建图像编码模型，另一个是寻找高效的熵编码方法。对于第一个问题，我们提出了一种新的图像编码模型，该模型使用基于空间滤波过程的分层结构。通过采用一定的技术对更高层的数据包进行保护，可以在一定程度上保证图像质量。对于第二个问题，我们提出了一种动态算术编码方法。在我们的方法中，少符号的估计概率是算术编码的参数之一。是从某些2的幂和的值中选择的。从静态和动态编码特性两方面确定了编码参数的转换方式。仿真结果表明，该方法与熵编码方法相比是有效的。最后，提出了基于分层编码方法的新型运动图像通信系统。

The Asynchronous Transfer Mode(ATM)is one of best methods as information transfer scheme because of its variety and availability. Over the ATM network, various types of communication connections can be realized and the network resources will be effectively used by statistical multiplexing. In moving image communication system using ATM, users can be access the network by unconstrained bit rate. so that the flame work of the coding scheme must be fundamentally changed, such that turning from constant bit rate to constant image quality should be planned.In this research, we decompose highly efficient image coding problem to two functional blocks, one is to construct coding models of images and another is to find the efficient entropy coding method.For the first question, we propose a new image coding model using a hierarchical structure based on a spatial filtering process. It can be keep picture quality to some extent by employing some technique to protect higher layer packets.For the second question, we propose a dynamic arithmetic coding method. In our method, the estimation probability of less symbol, which is one of parameters in arithmetic coding. is chosen from the values of some summations of powers of 2. The transition manner of the coding parameter is determined from both static and dynamic coding characteristics. Simulation results show that our proposed method is effective to the entropy coding method.Finally, we indicates the new types of moving image communication system based on the hierarchical coding method.