International Core Journal of Engineering 2020-26 | Page 172

TABLE I. P ERFORMANCE OF THE FIRST EXPERIMENTAL PLAN
All Intra Main10
Over VTM-5.0
Fig. 2. The first experimental plan.
Y U V EncT DecT
Class A1 -0.16% 0.08% 0.07% 121% 165%
Class A2 0.01% 0.16% 0.18% 117% 128%
Class B -0.05% 0.22% 0.22% 114% 140%
Class C -0.18% 0.05% 0.11% 125% 136%
Class E -0.05% 0.17% 0.21% 117% 128%
Overall -0.09% 0.14% 0.16% 119% 139%
Class D -0.18% 0.07% 0.10% 119% 127%
Class F -0.40% -0.06% 0.02% 108% 106%
TABLE II. P ERFORMANCE OF THE SECOND EXPERIMENTAL PLAN
All Intra Main10
Over VTM-5.0
Fig. 3. The second experimental plan.
IV. E XPERIMENTAL RESULT
The proposed method has been tested on the VTM5.0
reference software according to common test conditions
defined in JVET-N1010 [12]. The anchor is VTM5.0 with
implicit MTS enabled and the tests are proposed two
experimental plans. The following results were obtained as
shown table 1 and 2. The quantization parameters tested are
22, 27, 32 and 37. The all intra main10 configuration is used
as the test case. The Bjontegaard Delts-rate (BD-rate) [13-14]
is used to calculate the coding gain in the experiments. To
evaluate the encoding time and decoding time complexity of
the proposed method and anchor, we can use the following
formula to compute:
EncT
DecT
T pro
T anc
T pro
T anc
u 100% (1)
u 100% (2)
Y U V EncT DecT
Class A1 -0.39% -0.11% -0.19% 101% 92%
Class A2 -0.12% -0.03% -0.01% 100% 98%
Class B -0.12% 0.01% 0.01% 100% 105%
Class C -0.07% 0.09% 0.06% 105% 109%
Class E -0.17% -0.12% -0.03% 101% 104%
Overall -0.16% -0.02% -0.02% 101% 102%
Class D -0.05% -0.06% 0.14% 104% 107%
Class F -0.06% 0.01% 0.22% 98% 94%
V. C ONCLUSION
This paper analyzed that the difference between implicit
MTS and explicit MTS will affect the performance of
LFNST, and concluded that the combination of DCT2 and
DST7 of primary transform pair is the main factor that
affects the performance of LFNST, when implicit MTS is
enabled. Thus, we proposed to turn DST7 to DCT2 if
primary transform pair is a combination of DCT2 and DST7,
when implicit MTS is enabled. In order to verify the
reliability of the method, two experiments are designed to be
tested on VTM 5.0. Compared with the implicit MTS design
in VTM5.0, the experimental results show that the proposed
two method achieved 0.09% and 0.16% luma BD-rate
reduction for all intra configurations respectively, and almost
no influence on the coding complexity.
R EFERENCES
where T pro is the coding time of proposed and T anc is the
coding time of anchor. Table I shows the performance of the
proposed scheme one. From table I, we can see that the
scheme one achieved -0.09%, 0.14% and 0.16% coding gain
in Y, U and V component for all intra configurations
respectively, and the encoding and the decoding running time
complexities are about 119% and 139% respectively. [1]
Table II shows the performance of the proposed scheme
two. From table II, we can see that the scheme two achieved
-0.16%, -0.02% and -0.02% coding gain in Y, U and V
component for all intra configurations respectively, and the
encoding and the decoding running time complexities are
about 101% and 102% respectively. [3]
[2]
[4]
[5]
150
T. Wiegand, G. J. Sullivan, G. Bjontegaard, and A. Luthra,
“Overview of the H. 264/AVC video coding standard,” IEEE
Transactions on circuits and systems for video technology, vol. 13, no.
7, pp. 560-576, 2003.
G. J. Sullivan, J. Ohm, W.-J. Han, and T. Wiegand, “Overview of the
high efficiency video coding (hevc) standard,” IEEE Transactions on
Circuits and Systems for Video Technology, vol. 22, no. 12, pp.
1649-1668, 2012.
R. DONY et al., “Karhunen-loeve transform,” The transform and
data compression handbook, vol. 1, pp. 1-34: 29, 2001.
X. Zhao, J. Chen, M. Karczewicz, L. Zhang, X. Li, and W. J. Chien,
“Enhanced Multiple Transform for Video Coding,” in 2016 Data
Compression Conference (DCC), March 2016, pp. 73-82.
Jianle Chen, Yan Ye, and Seung Hwan Kim, “Algorithm description
for Versatile Video Coding and Test Model 2 (VTM 2)”, JVET-
K1002, Ljubljana,Macao , July 2018.