Abstract

The latest High-Efficiency Video Coding (HEVC) standard achieves nearly 50% bit rates reduction for similar quality relative to H.264/Advanced Video Coding(AVC) . However, its complexity is enormously increased ,which becomes one of the most challenges for its deployment in real time applications. The only solution to decrease the coding complexity is to set up different settings by adjusting various coding parameters. Among them, low complexity settings are suitable for industrial applications and conducive to the popularization of HEVC. Traditional fast mode decision algorithms mainly aim at decreasing coding complexity for high complexity settings. In this paper, we propose a fast mode decision method for HEVC with low complexity settings according to machine learning. A decision tree is constructed to decide whether to check 2N×2N mode or the SKIP/MERGE mode by exploiting relevant information from spatiotemporal adjacent Coding Units(CUs). Further mode skipping is performed based on the result of the first step. Experiments show that the proposed scheme can only increase by 1.42% Bjotegaard Delta Bit rate(BDBR) with an average time reduction of 22.45% for HEVC with low complexity settings.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call