4.5 Article

Fast HEVC Intra Mode Decision Based on RDO Cost Prediction

Journal

IEEE TRANSACTIONS ON BROADCASTING
Volume 65, Issue 1, Pages 109-122

Publisher

IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TBC.2018.2847464

Keywords

High efficiency video coding (HEVC); H.265; video compression; intra video coding; mode decision

Funding

  1. Vantrix Corporation
  2. Natural Sciences and Engineering Research Council of Canada through the Collaborative Research and Development Program [NSERC-CRD 428942-11]

Ask authors/readers for more resources

High efficiency video coding (HEVC) increases the number of intra coding modes to 35 to provide higher coding efficiency than previous video coding standards. This results in an increased encoder complexity, since there are more modes to be processed by the high resource-demanding rate-distortion optimization (RDO). In this paper, we propose a novel method to reduce the HEVC intra mode decision computational complexity and encoding time. This method is based on the prediction of the RDO cost of intra modes from a low-complexity sum of absolute transformed differences-based cost. By predicting the RDO cost, we are able to exclude non-promising modes from further processing and thereby save substantial computations. Also, a gradient-based method, using the Prewitt operator, is proposed to eliminate the non-relevant directional modes from the list of candidates. For even more complexity reduction, a mode classification is proposed to adaptively reduce chroma intra modes based on block texture. Experimental results show that we achieve a 47.3% encoding time reduction on average with a negligible quality loss of 0.062 dB for the Bjontegaard delta peak signal-to-noise ratio when we compare our method to the HEVC test model 15.0.

Authors

I am an author on this paper
Click your name to claim this paper and add it to your profile.

Reviews

Primary Rating

4.5
Not enough ratings

Secondary Ratings

Novelty
-
Significance
-
Scientific rigor
-
Rate this paper

Recommended

No Data Available
No Data Available