| High definition video and super high definition video bring about an ultimate immersion for human beings. It also presents a new challenge for video coding technology. Video coding is the key technology of “Internet Plus” and other practical video services. Massive video data compression is important for video storage and transmission. In 2013, the latest generation video coding standard – high efficiency video coding was proposed. HEVC is the inheritor of advanced video coding standard due to its reasonable coding framework. However the contradiction between complexity video coding and low bitrate video coding is the main obstacle that prevents the popularization of HEVC. This thesis focuses on theoretical study and practical application by achieving low complexity and low bitrate video coding.The major techniques and contributions of this thesis are listed as follows:1. This part is an overview of HEVC. Introduction is described as video coding standard development, HEVC extension standards and HEVC key techniques. Introduction is described as HEVC test model, including of coding structure and coding module. This part provides a primary knowledge of HEVC standard for beginner.2. This part proposes a low complexity video coding algorithm based on quadtree probability prediction mechanism. It costs a lot of encoding time that coding unit dividing process by traversing a complete quadtree. This part explores the relationship between coding unit statistical probability distribution and other parameters like quantization parameter, group of picture and content change. Based on this motivation, this part designs a quadtree probability prediction model and a method for regular model update. Experimental results show that the proposed low complexity video coding algorithm reduces almost 39% encoding time compared with HM15.0. The proposed algorithm is able to skip the unnecessary quadtree nodes while maintaining the accuracy of coding unit partitioning.3. This part proposes a low bitrate video coding algorithm based on difference frame. I frame is important for the quality of reconstructed image, but encoding I frame can cause bitrate increasing in the encoder output. Therefore it is a new idea that achieving low bitrate video coding by decreasing the bit cost of I frame. This part is based on information entropy theory and explores the multi-feature of the difference frame which can be used for I frame low bitrate encoding compatibility. In the end, this part raises the proposed low bitrate video coding algorithm based on difference frame. Experimental results show that the proposed low bitrate video coding algorithm reduces average 15% output bitrate compared with HM10.0. The proposed algorithm promotes the compression performance of HEVC and relieves the contradiction between reconstructed image quality and encoding recourse consumption. |
PDF Full Text Request