Scalable Video Coding (SVC) has been approved as the extension of the H.264/AVC video coding standard. This extension allows multiple frame rates, multiple resolutions, or multiple SNR rates of an image sequence to be contained in a single bit stream. The support of multiple resolutions within a single compressed bit stream is referred to as spatial SVC. It has achieved significant improvements in the enhancement layer coding efficiency due to the introduction of new modes and additional inter-layer prediction tools. However, the computational complexity of spatial SVC encoder is rather high, which inhibits it from practical use. Thus how to reduce the computational complexity but at the same time preserving the coding efficiency is quite crucial....[ Read more ]

Scalable Video Coding (SVC) has been approved as the extension of the H.264/AVC video coding standard. This extension allows multiple frame rates, multiple resolutions, or multiple SNR rates of an image sequence to be contained in a single bit stream. The support of multiple resolutions within a single compressed bit stream is referred to as spatial SVC. It has achieved significant improvements in the enhancement layer coding efficiency due to the introduction of new modes and additional inter-layer prediction tools. However, the computational complexity of spatial SVC encoder is rather high, which inhibits it from practical use. Thus how to reduce the computational complexity but at the same time preserving the coding efficiency is quite crucial.

In this thesis, three different methods to reduce the coding time are investigated and proposed: 1) reduce the use of inter-layer residual prediction tool; 2) reduce part of the rate-distortion cost calculations for each inter mode; 3) use a fast mode decision algorithm to accelerate the mode selection process. The three methods can be embedded into SVC's reference software-JSVM and used individually or be combined together. The simulation results demonstrate that the coding efficiency can be preserved well while consistent time saving over a large bit rate range can be achieved, especially when three methods are combined, over 50% time saving can be obtained.