一个高可尺度化视频编码方法的框架

A Highly Scalable Video Coding Framework

提出了一个高可尺度化的视频编码方法的框架。此框架采用三维子带变换对图象进行分解，以实现空间分辨率可尺度化和帧频可尺度化，将各子带数据进行分层量化以实现码率和失真度可尺度化。采用此方法的码流不仅可以同时支持分辨率、码率和失真度的可尺度化，而且可尺度化的粒度可以很细。

The widely used PIT(Progressive Image Transmission) can be progressive in resolution, in fidelity, or in both. Existing PIT methods that are progressive in both resolution and fidelity are quite complicated and difficult to implement. We present a highly scalable video coding framework that can make such PIT much simplified and easy to implement. Besides, this framework brings other advantages. In section 1, we used three dimensional subband transformation to achieve spatial resolution scalability and frame frequency scalability. In section 2, we used layered quantization of subband data to achieve rate scalability and distortion scalability. In section 3, we proposed a structure of rate stream to guarantee steady flow of data.