Direction-adaptive fixed length discrete cosine transform framework for efficient H.264/AVC video coding |
| |
| Affiliation: | 1. Department of Copyright Protection, Sangmyung University, Seoul, Korea;2. Department of Contents and Copyright, Sangmyung University, Seoul, Korea;1. College of Information Engineering, Capital Normal University, Beijing 100048, China;2. Beijing Advanced Innovation Center for Imaging Technology, Capital Normal University, Beijing 100048, China;3. Beijing Key Laboratory of Electronic System Reliability Technology, Capital Normal University, Beijing 100048, China;4. Queensland University of Technology, Brisbane QLD4000, Australia;5. Laboratory of Image Science and Technology, the Key Laboratory of Computer Network and Information Int egration, Ministry of Education, Southeast University, Nanjing 210096, China;6. Institut Mines -TELECOM, TELECOM Bretagne, INSERM 1101 LaTIM, France, Brest 29238, France;1. Institute of Digital Media and Communications, Taiyuan University of Science and Technology, 66 Waliu Road, Taiyuan, Shanxi, China;2. Institute of Image Processing, University of Electronic Science and Technology of China, 2006 Xiyuan Avenue, Hi-Tech (West) Zone, Chengdu, Sichuan, China |
| |
| Abstract: | The 2D-discrete cosine transform (2D-DCT) is one of the popular transformation for video coding. Yet, 2D-DCT may not be able to efficiently represent video data with fewer coefficients for oblique featured blocks. To further improve the compression gain for such oblique featured video data, this paper presents a directional transform framework based on direction-adaptive fixed length discrete cosine transform (DAFL-DCT) for intra-, and inter-frame. The proposed framework selects the best suitable transform mode from eight proposed directional transform modes for each block, and modified zigzag scanning pattern rearranges these transformed coefficients into a 1D-array, suitable for entropy encoding. The proposed scheme is analysed on JM 18.6 of H.264/AVC platform. Performance comparisons have been made with respect to rate-distortion (RD), Bjontegaard metrics, encoding time etc. The proposed transform scheme outperforms the conventional 2D-DCT and other state-of-art techniques in terms of compression gain and subjective quality. |
| |
| Keywords: | Discrete cosine transform Direction adaptive transform Directional transform framework Video coding Video compression |
| 本文献已被 ScienceDirect 等数据库收录！ |
|
