Clustered entropy computing: a holoentropy based encoding scheme for high efficient computing systems

This paper proposes a novel clustered entropy encoding scheme for high efficient computing systems for distributed video coding. As an exemplary system, high-efficiency video coding (HEVC) standard has been considered in this paper. The encoding process in HEVC system is performed using traditional entropy formula. While processing high resolution video sequences, it requires substantial improvement. To meet the requirement, this paper introduces clustered entropy computing, which distinguishes the video information has useful outliers. The pixel variations under varying frames are clustered based on the interestingness and the outliers are removed using an advanced entropy principle called as holoentropy. The proposed encoding scheme is adopted in the HEVC and the simulations are carried out. Experiments are conducted using eight benchmark video sequences and the statistical metrics such as mean, median, min, max and deviation measures are considered for investigating the PSNR versus compression ratio. In addition to this, the computational time is calculated for the best selection of the method. The experimental results show that the quality of the video is preserved more than the conventional encoding for increased compression ratios.