Scalable video encoding (SVC) block-level interlayer intra prediction (ILIP) method

Abstract

The invention discloses a scalable video encoding (SVC) block-level interlayer intra prediction (ILIP) method, which comprises the following steps of: at a coding end, sequentially calculating the variance of each block in a current coding frame of an enhancement layer (EL), and setting a filter control flag bit according to the variance values; if the flag bit is 1, calculating a self-correlation value among ILIP signals and a cross-correlation value between original EL signals and the ILIP signals; if all blocks in the frame are finished circulating, calculating a Wiener filter coefficient corresponding to the current frame; carrying out filtering operation on an ILIP result of the block, of which the filter control flag bit is 1 in the current coding frame of the EL by using the calculated Wiener filter coefficient, and further, judging whether the filtering should be carried out on the ILIP result of the block or not according to a filter result; and coding the filter control flagbit and the Wiener filter coefficient, then, writing in a header message, and transmitting the header message to a decoding end. With the adoption of the method disclosed by the invention, the accuracy of the ILIP technology can be effectively enhanced, and the coding efficiency of a coder is increased.

Description

technical field [0001] The invention relates to video compression technology, in particular to a block-level inter-layer intra-frame prediction method of scalable video coding in the video compression technology. Background technique [0002] As an extension of the H.264 / AVC video coding standard, Scalable Video Coding (Scalable Video Coding, SVC) provides three classification methods in the time domain, space domain and quality domain to solve the heterogeneous transmission and storage systems in video communication question. Among them, the time domain grading method can be realized by hierarchical B frame (Hierarichal B) technology; the spatial domain grading and quality domain grading are realized by using a layered coding structure - the base layer (Base Layer, BL) coded frame adopts the traditional H.264 / AVC encoding method has the smallest spatial resolution and poor quality; the enhanced layer (Enhanced Layer, EL) encoded frame has greater spatial resolution and hi...

AI Technical Summary

Problems solved by technology

However, if the Wiener filter coefficient training and filtering operation are performed in units of one frame, the filter control is not precise enough, so there is still room for further improvement in the accuracy of the ILIP signal in SVC

[0007] In summary, the results of ILIP in SVC are not accurate enough, and the performance of existing optimization algorithms needs to be further improved.

Images