| 摘要: | 研究期間：10108~10207;As much higher definition videos are required in today’s multimedia service, the ITU-T VECG and ISO/IEC MPEG groups have been recently investigating the next-generation video coding standard: the high efficiency video coding standard (HEVC, HM4.0 reference software). The new proposed HEVC coding technique features many advanced techniques, such as larger prediction sizes, more intra prediction modes, merge coding technique, novel in-loop filter technique…etc. These new advanced techniques employed in HEVC lead to a higher coding performance, compared to H.264/AVC. For example, the HEVC is proposed to achieve up to 50% of bit-rate reduction at the same visual quality, when compared with H.264/AVC, but with extremely high computational complexity. To meet high computation requirement for HEVC, in this three-year’s project, we will investigate the HEVC coding technique in the cloud computing systems that provides very robust computation abilities. We will investigate an advanced HEVC coding technique that employs the down-sampling and prediction techniques to further improve the coding performance of HEVC. In the proposed advanced HEVC, we consider four different HEVC coding modes (full frame mode, vertical down-sampling mode, horizontal down-sampling mode and vertical&horizontal down-sampling mode respectively) and choose the best mode for coding each frame or each tree-block based on rate distortion optimization (RDO) function. In the first year, we will investigate the advanced intra frame coding technique for HEVC. In the second year, we will focus on the advanced inter frame scheme for HEVC. In the last year, the computation resource allocations for the four modes in cloud computing systems will be studied and integrated to significantly reduce the overall computational complexity. We will finally concentrate on the integration of this sub-project with other sub-projects, and provide an efficient high-quality video service system in the cloud.As much higher definition videos are required in today’s multimedia service, the ITU-T VECG and ISO/IEC MPEG groups have been recently investigating the next-generation video coding standard: the high efficiency video coding standard (HEVC, HM4.0 reference software). The new proposed HEVC coding technique features many advanced techniques, such as larger prediction sizes, more intra prediction modes, merge coding technique, novel in-loop filter technique…etc. These new advanced techniques employed in HEVC lead to a higher coding performance, compared to H.264/AVC. For example, the HEVC is proposed to achieve up to 50% of bit-rate reduction at the same visual quality, when compared with H.264/AVC, but with extremely high computational complexity. To meet high computation requirement for HEVC, in this three-year’s project, we will investigate the HEVC coding technique in the cloud computing systems that provides very robust computation abilities. We will investigate an advanced HEVC coding technique that employs the down-sampling and prediction techniques to further improve the coding performance of HEVC. In the proposed advanced HEVC, we consider four different HEVC coding modes (full frame mode, vertical down-sampling mode, horizontal down-sampling mode and vertical&horizontal down-sampling mode respectively) and choose the best mode for coding each frame or each tree-block based on rate distortion optimization (RDO) function. In the first year, we will investigate the advanced intra frame coding technique for HEVC. In the second year, we will focus on the advanced inter frame scheme for HEVC. In the last year, the computation resource allocations for the four modes in cloud computing systems will be studied and integrated to significantly reduce the overall computational complexity. We will finally concentrate on the integration of this sub-project with other sub-projects, and provide an efficient high-quality video service system in the cloud. |
