Conference Paper

A configurable and programmable motion estimation processor for the H.264 video codec

Electron. Eng. Dept., Bristol Univ., Bristol
DOI: 10.1109/FPL.2008.4629923 Conference: Field Programmable Logic and Applications, 2008. FPL 2008. International Conference on
Source: IEEE Xplore

ABSTRACT This work presents a programmable, configurable motion estimation processor for the H.264 video coding standard, capable of handling the processing requirements of high definition (HD) video and suitable for FPGA implementation. The programmable aspect of the processor follows the ASIP (application specific instruction set processor) approach with a instruction set targeted to accelerating block matching motion estimation algorithms. Configurability relates to the ability to optimize the microarchitecture for the selected algorithm and performance requirements through varying the number and type of execution units at compile time.

  • [Show abstract] [Hide abstract]
    ABSTRACT: This paper proposes Integer-pel Motion Estimation (IME) specific instructions and their hardware architecture for Application Specific Instruction-set Processor (ASIP). With parallel SAD Processing Elements (PEs) using pattern information, the proposed IME instruction supports not only the full search algorithm but also various fast search algorithms. Moreover, the revisiting prevention technique enables that the proposed ASIP can efficiently perform the fast search operations. The gate count is 43K gates for each Processing Element Group (PEG) which has 256 SAD PEs. The proposed ASIP with eight PEGs runs at 160MHz and can handle 1080p@30 frames in real-time. I. INTRODUCTION Video coding techniques provide efficient solutions to represent video data in a more compact and robust way. To meet the industry requirements of video techniques, various standards, such as MPEG-4, H.264/AVC (1), High- performance Video Coding (2), etc. have been developed. The recent video coding technologies adopt various sophisticated tools to refine coding process like integer discrete cosine transform, in-loop filter, variable length coding, intra prediction, motion estimation (ME), etc. Among these tools, ME is the most important and complex part of the video codec. ME algorithms are usually implemented as Application- specific Integrated Circuit (ASIC). Although ASIC-based designs have high performance, low power consumption and area minimization, its flexibility is exceptionally low. Thus, whenever standards and specifications change, ASIC might be redesigned and remanufactured and thus the Non-recurring Engineering (NRE) cost and time-to-market should inevitably increase. Hence, ASIP based designs have been suggested as alternatives. The ASIP approach has relatively more flexibility compared with the ASIC approach. Consequently, the design time and NRE cost could be reduced when ASIP is used. Existing DSPs have various application specific instructions for multimedia algorithms. However, the newly adopted features of H.264/AVC, conventional instructions are not efficient to support H.264/AVC. To efficiently implement the new features of H.264/AVC on ASIP, the novel instructions and their hardware architecture have been proposed in (3). Hence, this paper proposes the IME specific instructions for an ME ASIP, which include SAD, Compare, and Mode Decision instructions. We also design their efficient hardware architectures. The proposed ME ASIP can handle real-time processing for HD video and various motion estimation algorithms can be easily programmed.
    International Symposium on Circuits and Systems (ISCAS 2011), May 15-19 2011, Rio de Janeiro, Brazil; 01/2011
  • [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents a reconfigurable processor designed to execute user-defined block-matching motion estimation algorithms, and a toolset for the design of such algorithms and for the configuration of the processor. The toolset enables the exploration of the processor's design space in order to find an optimal configuration depending on the target application. The use of the toolset to test different configurations for different kinds of video sequences is illustrated. Experimental results show the benefits and cost of certain optimizations in the motion estimation process, and that fast block-matching search algorithms can outperform full search algorithms commonly used in hardware implementations. The usefulness of the toolset in exploring the configuration space is also shown.
    Field Programmable Logic and Applications, 2009. FPL 2009. International Conference on; 10/2009
  • [Show abstract] [Hide abstract]
    ABSTRACT: This work investigates how the dynamic reconfiguration features available in modern FPGAs can be combined with a voltage-frequency scaling strategy to adapt the processing performance available in the system to the available energy budget. As defined by a hardware operating system, each processing node can configure itself as a virtual processor able to execute a control algorithm that will adjust the voltage-frequency operational point of the reconfigurable fabric so that functions mapped to this fabric meet their processing deadlines. Once the new operational point has been reached the processor is removed from the fabric. This is then reconfigured to implement the application-specific logic exploiting the dynamic nature of FPGAs.
    Control Applications, (CCA) & Intelligent Control, (ISIC), 2009 IEEE; 08/2009