A new intra prediction method based on consistent luminance changes

作者:Cao X.; Zhu C.; Yang M.; Lin Y.; Zheng J.
来源:9th Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2017, 2017-12-12 To 2017-12-15.
DOI:10.1109/APSIPA.2017.8282088

摘要

Intra prediction is an effective way to remove spatial redundancy in video coding. Currently, most intra prediction methods are based on the locally reconstructed pixels to predict the current block. In this paper, we introduce a new intra prediction method that uses non-local information to conduct consistent luminance change prediction (LCP) for current block, where the least square error model is used to obtain the LCP parameters. Compared with the intra template matching prediction (TMP) method, which may introduce the mismatch between the reference template and the corresponding reconstruction block, the LCP method can mitigate the mismatch problem of TMP. The LCP method can achieve average bit-rate savings of 0.58% with acceptable increase in encoding time and decoding time. For high resolution (HR) video sequences, the algorithm can achieve up to 2.36% bit-rate savings.

全文