DC Coefficient Estimation of Intra-Predicted
Residuals in High Efficiency Video Coding
Chen Chen
#1
, Zexiang Miao
∗2
, Xiandong Meng
#3
, Shuyuan Zhu
∗4
, Bing Zeng
∗5
#
Department of Electronic & Computer Engineering, The Hong Kong University of Science and Technology
Clear Water Bay, Kowloon, Hong Kong, China
1
cchenaf@ust.hk;
3
xmengab@ust.hk;
∗
Institute of Image Processing, University of Electronic Science and Technology of China
No.2006, Xiyuan Avenue, Chengdu, Sichuan, China
2
zxmiao@std.uestc.edu.cn;
4
eezsy@uestc.edu.cn;
5
eezeng@uestc.edu.cn
Abstract—This paper proposes a DC coefficient estimation
algorithm for intra-predicted residual blocks in the High Ef-
ficiency Video Coding (HEVC) standard. Discarding the DC
coefficient in the current coding block leads to a substantial
bit-saving but produces at the same time strong discontinuities
between this block and its neighboring reconstructed blocks. To
overcome this problem, we propose an estimation algorithm for
the DC coefficient, which solves an optimal offset in a closed-
form in the pixel domain to recover the corresponding block
edges. Test results show that our algorithm achieves 1.0% and
1.4% BD-rate reduction on average for luma and chroma as
compared with HM-16.6, respectively, when the sign-bit-hiding
(SBH) technique is disabled. When SDH is set on, namely under
the common test condition (CTC), the BD-rate reduction drops
slightly to 0.7% and 1.1% for luma and chroma, respectively. In
the meantime, the test results show that both encoding time and
decoding time increase only slightly (about 10%, without any
special optimization on programming our proposed algorithm).
Index Terms—HEVC, video coding, intra prediction, coefficient
estimation, R-D performance
I. INTRODUCTION
The High Efficiency Video Coding (HEVC) standard, also
known as H.265, was released in early 2013. Compared to its
predecessor - H.264/AVC, developed by the joint collaborative
team on video coding (JCT-VC), HEVC achieves 50% bit-rate
reduction under the same perceptual quality [1].
While inheriting most of H.264/AVC’s intra coding features,
HEVC introduces a large number of new and advanced coding
techniques. For instance, it expands H.264/AVC’s 9 intra
prediction modes to as many as 35 (including 33 angular
modes, the DC mode, and the planar mode) for all block sizes
on luma coding; its quad-tree structure allows a 64×64 coding
tree unit (CTU) to be split down to 8×8 coding unit (CU); and
a large scope of self-adaptive size selection for prediction unit
(PU) and transform unit (TU) is supported. However, HEVC
is still based on a hybrid coding scheme using the block-based
prediction, transform, quantization, and entropy coding.
At the quantization step, HEVC introduces a new technique
for coding the transform coefficients, called sign bit hiding
(SBH) [2], [3]. Depending on some pre-defined criterion, the
sign value of the first non-zero coefficient along the scan order
in a coding group (CG) is simply hidden into the parity of the
sum of the levels within the CG: even corresponding to “+”
(positive) and odd to “-” (negative) so that one coefficient in
a CG perhaps needs to be fine tuned to satisfy the convention.
In past few years, people have been dedicated in recovering
transform coefficients corrupted by channel noise [4], [5]. But
such methods can only be employed in decoding terminal
for post processing which is useless to improve compression
efficiency. On the other hand, several inpainting algorithms
have been used to recover the missing bits or coefficients
[6]. However, these inpainting algorithms are usually too
complicated to be implemented in real-time, particularly at the
decoder side. Consequently, we are inspired to design a much
more efficient estimation algorithm in which the directional
texture information of the neighboring blocks will be utilized
and the estimated DC value can be calculated in a closed-form.
In this paper, we propose to estimate the DC coefficient
of the current transform block from its neighboring blocks
that have already been coded (i.e., above or to the left of the
current block). Instead of coding the original DC coefficient,
we code the difference between the original DC coefficient
and the estimated one, which may lead to a bit-saving if the
estimation is accurate. The rest of the paper is organized as
follows. In Section II, we present the initial idea of performing
an estimation of the DC coefficient, which is largely motivated
by some test results of HEVC (in terms of the bit-saving) when
the DC coefficients of all coding blocks are discarded. Then,
we describe in detail in Section III how to design an efficient
estimation method for the DC coefficient of the current block.
Some experimental results are given in Section IV to verify
the performance of our proposed algorithm. Finally, some
conclusions and discussions are presented in Section V.
II. INITIAL IDEA AND MOTIVATIONS
Figure 1 sketches the procedure of our initial idea for
performing the DC coefficient estimation. Specifically, we
apply the same HEVC transform and quantization on predict-
ed residuals that are obtained by subtracting the prediction
block by DC mode from the original one. The top-right
image is reconstructed following the regular de-quantization
and inverse transform and the bottom-right image is recon-
structed without the DC coefficient—denoted by E
0,0
(i.e.,
978-1-5090-5316-2/16/$31.00
c
2016 IEEE VCIP 2016, Nov. 27 – 30, 2016, Chengdu, China
资源评论
