| Scalable Video Coding (SVC) is an advanced video coding technology. Its main idea isto produce a basic layer which contains basic picture quality, and one or more enhance layersthat contain additional information. Enhance layer is dependent of basic layer. The decodercould decode the basic layer.However, the enhance layer needs the data of basic layer fordecoding in order to produce the enhanced picture. For systems that implement SVC, theencoder would send basic layer data to end users who have bad channel condition. As to thosewho obtain higher bandwidth, the encoder would send both basic and enhance layer toachieve better quality. According to different coding method, enhance layer could improvethe basic layer from time resolution, space resolution, space-frequency resolution and SNR.This dissertation first presents the history of SVC, especially the research achievementsduring recent years. It also provides software implementation of several prevalent codingmethods, together with experiments and analysis of the results.Secondly, with experiment results it's found that the high bit rate of enhance layer hasconstraints on broad application of SVC. Analyzing several improvements on SVC, it's alsofound that research on using ROI to enhance layer is less focused. Therefore, this dissertationtakes in-depth research on ROI theory, designs SFC, FIF and LIF to extract ROI, andoptimizes these functions according to effects.Finally, this dissertation applies those functions, using ROI theory to guide SVC. It doesexperiments on space SVC and SNR SVC. With results it could see that using ROI in SVCcould greatly reduce the enhance layer bit rate, however, the details could be preserved. |
