A Fast Mode Decision Algorithm and Its Hardware Design for H.264/AVC Intra Prediction - Advances in Image and Graphics Technologies - page 48

A Fast Mode Decision Algorithm and Its Hardware
Design for H.264/AVC Intra Prediction

Wei Wang 1 , Yuting Xie, Tao Lin, and Jie Hu

1 College of Electronics Engineering, Chongqing University of Posts and Telecommunications,
Chongqing 400065, China
[e-mail]

Abstract. This paper presents an architecture of mode decision algorithm for
H.264/AVC intra prediction. In the algorithm design, based on the inherent cor-
relation existing in the spatial prediction modes, a significant computational
savings can be achieved. In the hardware design, through efficient sharing of
configurable units and parallel executions of different candidate prediction
modes, a lower hardware utilization and higher execution speed can be
achieved. Synthesis results show that the proposed architecture can process
HDTV (1920×1080) video at 60 fps in FPGA platform and maximum frequen-
cy achieved is 184.8 MHz.

Keywords: H.264/AVC, intra prediction, spatial correlation, hardware verifica-
tion and FPGA.

1 Introduction

With the wide use of technologies such as online video, digital television and video
conference, video compression technology has become an inevitable component for
storage and transmission. Currently, H.264/AVC is published as the new generate
video compression coding standard which achieves the highest compression perfor-
mance without sacrificing the quality of picture [1].
The high video compression efficiency of the H.264/AVC standard is achieved
through a particular combination of a number of encoding tools but rather any single
feature. Intra prediction algorithm is the important part of the encoder which occupies
the most calculation and generates a prediction for a macroblock based on the spatial
redundancy. However, the better compression efficiency comes with a high computa-
tion complexity which makes it different from meeting the demand of the real time [2].
The intra prediction algorithms with lower complexity and high-speed computing
is important for the real-time requirements. Huang et al. [3] proposed an algorithm
with better efficiency on mode decision, but it suffered from a higher computational
complexity in determining the partitioning. Yu et al. [4] put forward a fast algorithm
in determining the partitioning to reduce the high computational complexity, but theur
argument shows the insufficiency of efficiency on mode decision.
An efficient architecture for intra prediction is proposed based on the inherent
correlation existing in the intra prediction modes which can reduce the candidate