Macroblock Mode Pre-classification Algorithms
based on Motion Vectors Filtering for
H.264/AVC
Cheng Zhou
College of Electronic Information Engineering, South-Center University for Nationalities, Wuhan, China
Email:zcresearch@gmail.com
Yanhong Lu
Wuhan Polytechnic, Wuhan, China
Email: luyanhong126@gmail.com
Chengyi Xiong
Hubei Key Laboratory of Intelligent Wireless Communications, South-Center University for Nationalities, Wuhan,
China
Email: xiongcy@mail.scuec.edu.cn
Zhirong Gao
College of Computer Science, South-Center University for Nationalities, Wuhan, China
Email: gaozhirong@mail.scuec.edu.cn
Abstract—Noise robustness of video codec is a big challenge.
This paper proposes a novel macroblock mode pre-
classification algorithm based on the co-matching criteria,
motion vectors spatial and temporal filtering. The proposed
algorithm could distinguish the moving object from the
background noise. Simulations show that this approach can
result in a time savings of over 62.86% for several typical
sequences with noise. It also reduces the average
Bjontegaard delta bit rate by about 1.67%, and increases
the average Bjontegaard delta peak signal-to-noise ratio by
about 0.08dB, compared with the algorithm of H.264/AVC.
Index Terms—video coding, noise robustness, motion
estimation, mode decision, motion vectors filtering,
H.264/AVC
I.
I
NTRODUCTION
With the rapid growth of network video system,
compression and pre-processing analysis of massive
digital video becomes more and more important. High
efficient video coding technology is the key to solve these
problems. However, it appears many problems in the
practical application of video system. Firstly, when video
noise increases, the video coding efficiency considerably
decreases. Secondly, the new generation of video coding
standard such as H.264/AVC [1] can obtain higher
compression efficiency [2]. Unfortunately, it is difficult
to design the hardware codec because of the complex
coding control model [3]. Thirdly, as the complexity of
the new generation of video coding standard is very high,
the pre-processing and analysis of video data become
more and more difficult [4].
This paper focuses on how to increase the noise
robustness of video encoder. And the main contribution
of this thesis is a novel macroblock mode pre-
classification method based on co-matching criteria,
motion vectors spatial and temporal filtering. The method
uses the co-matching criteria to judge the current
macroblock in order to eliminate the noise impact and use
the temporal and spatial filtering of the motion vector
fields of encoded frames to eliminate the noise motion
vectors. According to the motion information of current
macroblock, the method limits the coding mode of
current macroblock. The proposed algorithm can improve
the noise robustness of the H.264/AVC encoder.
Simulations show that this approach can result in a time
savings of over 62.86% for several typical sequences with
noise.It also reduces the average Bjontegaard delta bit
rate by about 1.67%, and increases the average
Bjontegaard delta peak signal-to-noise ratio by about
0.08dB, compared with the algorithm of H.264/AVC.
Experiments prove that this algorithm improves the
coding performance and coding speed.
II.
R
ELATED WORK
A. Coding Control Model in H.264/AVC
The H.264/AVC standard has been widely applied to
real-time video system because it offers significantly
better coding efficiency than previous standards [5]. The
improvements are mainly obtained from various state-of-
the-art techniques such as intra predictions, variable
block-size motion compensation and quarter-pixel motion
This work is supported by
National Natural Science Foundation of
China under Grant # 61201268 to Cheng Zhou
Science Foundation of China under Grant # 2010CDZ022 to Cheng
Zhou.
C
orresponding author: Yanhong Lu
JOURNAL OF MULTIMEDIA, VOL. 8, NO. 2, APRIL 2013 129
© 2013 ACADEMY
doi:10.4304/jmm.8.2.129-136