摘要
The growing popularity of Internet applications and services has rendered high subjective video quality crucial to the user experience. Increasing needs for better video resolution and faster transmission bandwidths present challenges to the goal of achieving balance between video quality and coding cost. In this paper, we propose a Perceptive Variable Bit-Rate Control （PVBRC） framework for the state-of-the-art video coding standard High-Efficiency Video Coding （HEVC）/H.265. PVBRC allocates a bit-rate to a picture while taking a Comprehensive Picture Quality Assessment （CPQA） model and perceptive target bit-rate allocation into consideration. The CPQA model calculates the objective and perceptive quality of both source and reconstructed pictures by referring to the human vision system. The coding bit-rate is then dynamically allocated by the result of the CPQA model according to differences in picture content. In PVBRC, the quantization parameter for current picture encoding is updated by an effective fuzzy logical controller to satisfy the transmission requirements of the Internet of Things. Experimental results show that the proposed PVBRC can achieve average bit savings by 11.49% when compared with constant bit-rate control under the same objective and subjective video quality.
The growing popularity of Internet applications and services has rendered high subjective video quality crucial to the user experience. Increasing needs for better video resolution and faster transmission bandwidths present challenges to the goal of achieving balance between video quality and coding cost. In this paper, we propose a Perceptive Variable Bit-Rate Control （PVBRC） framework for the state-of-the-art video coding standard High-Efficiency Video Coding （HEVC）/H.265. PVBRC allocates a bit-rate to a picture while taking a Comprehensive Picture Quality Assessment （CPQA） model and perceptive target bit-rate allocation into consideration. The CPQA model calculates the objective and perceptive quality of both source and reconstructed pictures by referring to the human vision system. The coding bit-rate is then dynamically allocated by the result of the CPQA model according to differences in picture content. In PVBRC, the quantization parameter for current picture encoding is updated by an effective fuzzy logical controller to satisfy the transmission requirements of the Internet of Things. Experimental results show that the proposed PVBRC can achieve average bit savings by 11.49% when compared with constant bit-rate control under the same objective and subjective video quality.
基金
supported by Foundation of Science and Technology Department of Sichuan Province (Nos. 2017JY0007 and 2017HH0075)
