Abstract
To improve compression performance, High Efficiency Video Coding (HEVC) employs a quad-tree based block representation, namely Coding Tree Unit (CTU), which can support larger partitions and more coding modes than a traditional macroblock. Despite its high compression efficiency, the number of combinations of coding modes increases dramatically, which results in high computational complexity at the encoder. Here we propose a flexible framework for HEVC coding mode selection, with a user-defined global complexity factor. Based on linear programming, a hierarchical complexity allocation scheme is developed to allocate computational complexities among frames and Coding Units (CUs) to maximize the overall Rate-Distortion (RD) performance. In each CU, with the allocated complexity factor, a mode mapping based approach is employed for coding mode selection. Extensive experiments demonstrate that, with a series of global complexity factors, the proposed model can achieve good trade-offs between computational complexity and RD performance. © 2007-2012 IEEE.
Original language | English |
---|---|
Pages (from-to) | 1135-1144 |
Journal | IEEE Journal on Selected Topics in Signal Processing |
Volume | 7 |
Issue number | 6 |
Early online date | 27 Jun 2013 |
DOIs | |
Publication status | Published - Dec 2013 |
Externally published | Yes |
Keywords
- Complexity allocation
- HEVC
- linear programming
- mode decision
- video coding