Spatially Scalable Video Coding (SSVC) Using Motion Compensated Recursive Temporal Filtering (MCRTF)

摘要

Through the following years, streaming makers will be progressively tasked supplying enhanced streams of video to gadgets as mobile phones and set top boxes, alongside diverse quality variants for clients to get content on general Internet. While there have been various ways to deal with this issue, including different bit rate feature, one exceptionally solid competitor will be a H.264 expansion called Scalable Video Coding ( SVC). It encodes video into "layers," beginning with the "base" layer, which contains the most minimal information of the bit-stream, and then moving towards “enhanced layers” which includes the information to scale up the output. Also SVC gives support for different resolutions inside a single compressed bit stream which is known as spatial scalabilility. In this thesis a problem on SSVC has been addressed. The video sequences had been made scalable in spatial domain. In order to make it more efficient for real time applications, motion compensated recursive temporal filtering (MCRTF) has been implemented. This scheme enhances the efficiency of the components of a visual signal. The temporal filter used here helps in reducing noisearising from the plurality of the frames and the improvised output with reduced noise is used in the process of predictive encoding. Also it eliminates the inherent drift, which arises due to difference between encoder and decoder. As visual signals are always subjected to temporal correlation, motion compensation from the adjacent frames and using it as the reference during the process of predictive coding is of prior importance. The conventional and the proposed method have been used during the encoding process of various video sequences in the spatial domain and an analytical study on that has been carried out