Sample-based Weighted Prediction
for Lossless Enhancement Layer Coding
in HEVC
Andreas Heindel, Eugen Wige, André Kaup
Lehrstuhl für Multimediakommunikation und Signalverarbeitung
Friedrich-Alexander-Universität Erlangen-Nürnberg
Outline
●  Motivation
●  System Overview
●  Enhancement Layer Coding
●  Experimental Results
●  Conclusion
Heindel, Wige, Kaup: "SWP for Lossless EL Coding in HEVC"
Chair of Multimedia Communications and Signal Processing
Page 2
Jan. 30, 2014
Motivation
●  New video coding standard HEVC primarily targeting consumer
applications with lossy compression
●  Need for lossless compression in professional applications
-  Medical imaging (telemedicine)
-  Archiving (cinema)
-  …
●  High bitrate ⇔ limited channel capacity
⇒ Scalable lossless coding using two layers
-  Lossy base layer (BL)
-  Lossless enhancement layer (EL)
en.wikipedia.org/wiki/File:RupturedAAA.png
Heindel, Wige, Kaup: "SWP for Lossless EL Coding in HEVC"
Chair of Multimedia Communications and Signal Processing
Page 3
Jan. 30, 2014
System Overview
Heindel, Wige, Kaup: "SWP for Lossless EL Coding in HEVC"
Chair of Multimedia Communications and Signal Processing
Page 4
Jan. 30, 2014
Base Layer
●  Lossy BL compression using HEVC
Heindel, Wige, Kaup: "SWP for Lossless EL Coding in HEVC"
Chair of Multimedia Communications and Signal Processing
Page 5
Jan. 30, 2014
Enhancement Layer
●  Lossless EL coding using the proposed
Sample-based Weighted Prediction for Enhancement Layer Coding
(SELC)
Heindel, Wige, Kaup: "SWP for Lossless EL Coding in HEVC"
Chair of Multimedia Communications and Signal Processing
Page 6
Jan. 30, 2014
Enhancement Layer Coding
SELC Encoder
SELC Decoder
Intra prediction:
●  Sample-based Weighted Prediction (SWP) [1]
⇒ SWP for Enhancement Layer Coding (SELC)
Entropy coding/decoding:
●  CABAC
[1] E. Wige et al., “Pixel-based averaging predictor for HEVC lossless coding,” in Proc. of IEEE ICIP, Sep. 2013.
Heindel, Wige, Kaup: "SWP for Lossless EL Coding in HEVC"
Chair of Multimedia Communications and Signal Processing
Page 7
Jan. 30, 2014
Intra Prediction (SWP) I
●  Four-pixel neighborhood and four-pixel patch
Neighborhood of
current pixel
Patch around a pixel
Current pixel
Patch pixel
current pixel
●  Patch around the current pixel is compared
to the patches of the neighborhood pixels ...
(-1,-1)
Current pixel
shift=(0,0)
(0,-1)
(1,1)
(-1,0)
[3] P. Amon et al., “RCE2: Sample-based weighted intra prediction for lossless coding,” document JCTVC-M0052, JCT-VC, Apr. 2013.
Heindel, Wige, Kaup: "SWP for Lossless EL Coding in HEVC"
Chair of Multimedia Communications and Signal Processing
Page 8
Jan. 30, 2014
Intra Prediction (SWP) II
Algorithm
1.  Compute the Sum of Absolute Distances (SAD) between…
-  patch around the current pixel
-  patches around the pixels
of the current neighborhood
2.  Map SAD values to weights:
3.  Compute prediction for current pixel:
: neighborhood
[1] E. Wige et al., “Pixel-based averaging predictor for HEVC lossless coding,” in Proc. of IEEE ICIP, Sep. 2013.
Heindel, Wige, Kaup: "SWP for Lossless EL Coding in HEVC"
Chair of Multimedia Communications and Signal Processing
Page 9
Jan. 30, 2014
Entropy Coding (CABAC)
1.  Binarization of the prediction error values
abs_value_remaining:
•  Truncated Rice code
like with coeff_abs_level_remaining in HEVC [4]
•  Exponential Golomb code
•  Adaptive update of the Rice parameter
2.  Context modeling
- 
- 
Two models
One for b0 and another for b2
3.  Arithmetic Coding
[4] ITU-T Rec. H.265 and ISO/IEC FDIS 23008-2: High Efficiency Video Coding (HEVC), JCT-VC Std., Apr. 2013.
Heindel, Wige, Kaup: "SWP for Lossless EL Coding in HEVC"
Chair of Multimedia Communications and Signal Processing
Page 10
Jan. 30, 2014
Experimental Results
Coding efficiency:
Relative bitrate differences1 for EL coding compared to SHM-2.1
HM-11.0
SELC
QP22
QP27
QP32
QP37
QP22
QP27
QP32
QP37
1.2%
1.0%
0.3%
0.8%
-2.6%
-4.7%
-6.5%
-7.3%
Runtime:
Relative runtime increase2 for EL processing compared to BL processing only
SHM-2.1
1:
HM-11.0
SELC
QP22
QP27
QP32
QP37
QP22
QP27
QP32
QP37
QP22
QP27
QP32
QP37
Enc
25.3%
30.6%
34.9%
37.7%
18.5%
22.5%
25.5%
27.7%
0.6%
0.7%
0.9%
0.8%
Dec
244.4% 338.9% 443.6% 536.2% 260.1% 361.9% 451.4% 533.8% 202.8% 279.6% 334.3% 374.8%
average values w/o ElFuente
2:
average values for all sequences
Heindel, Wige, Kaup: "SWP for Lossless EL Coding in HEVC"
Chair of Multimedia Communications and Signal Processing
Page 11
Jan. 30, 2014
Conclusion
●  Method for lossless EL compression
SWP for Enhancement Layer Coding (SELC)
●  Two steps:
-  Intra prediction using SWP
-  Entropy coding using CABAC
●  Results
-  Up to 7.3% bitrate reduction compared to SHM-2.1 (average: 5.3%)
-  Negligible complexity increase in the encoder
⇒ Software EL processing on top of hardware BL coding possible
-  Considerably faster decoding than SHM-2.1 or HM-11.0
Heindel, Wige, Kaup: "SWP for Lossless EL Coding in HEVC"
Chair of Multimedia Communications and Signal Processing
Page 12
Jan. 30, 2014
References
[1] E. Wige, G. Yammine, P. Amon, A. Hutter, and A. Kaup, “Pixel-based averaging
predictor for HEVC lossless coding,” in Proc. of IEEE International Conference on
Image Processing (ICIP), Melbourne, Australia, Sep. 2013.
[2] P. Amon, A. Hutter, E. Wige, and A. Kaup, “Intra prediction for lossless coding,”
document JCTVC-L0161, ITU-T VCEG and ISO/IEC MPEG (JCT-VC), Geneva,
Switzerland, Jan. 2013.
[3] P. Amon, A. Hutter, E. Wige, and A. Kaup, “RCE2: Sample-based weighted intra
prediction for lossless coding,” document JCTVC-M0052, ITU-T VCEG and ISO/IEC
MPEG (JCT-VC), Incheon, Republic of Korea, Apr. 2013.
[4] ITU-T Rec. H.265 and ISO/IEC FDIS 23008-2: High Efficiency Video Coding (HEVC),
ITU-T VCEG and ISO/IEC MPEG (JCT-VC) Std., Apr. 2013.
Heindel, Wige, Kaup: "SWP for Lossless EL Coding in HEVC"
Chair of Multimedia Communications and Signal Processing
Page 13
Jan. 30, 2014
Special Case: ElFuente
…
…
…
…
…
…
…
…
…
Frame 1-75
•  Almost entirely black content of original and residual
•  HM, SHM:
•  All transform coefficient levels are 0
Frame 76-300
⇒  Coded block flags are set to 0
⇒  Efficient coding of whole blocks
•  SELC: Sample-wise processing
•  Solution: Check if BL is already lossless (e.g. block-wise)
Heindel, Wige, Kaup: "SWP for Lossless EL Coding in HEVC"
Chair of Multimedia Communications and Signal Processing
Page 14
Jan. 30, 2014
Intra Prediction (SWP)
Current pixel
(-1,-1)
Patch pixel
Example:
•  Current pixel
•  Neighborhood pixel
•  Here:
1.  Compute SAD values for all neighbors:
Set of shifts for the patch:
2.  Map SAD values to weights:
Precomputed LUT:
3. 
Compute prediction
Heindel, Wige, Kaup: "SWP for Lossless EL Coding in HEVC"
Chair of Multimedia Communications and Signal Processing
Page 15
Jan. 30, 2014
CABAC: Rice Parameter Update
●  Binarization of the prediction error values:
abs_value_remaining:
•  Truncated Rice code
•  Exponential Golomb code
●  Pixel-wise update of the Rice parameter:
●  Binarization of abs_value_remaining
-  Like coeff_abs_level_remaining in HEVC [4]
-  Rice parameter update extended to allow decrease of
[4] ITU-T Rec. H.265 and ISO/IEC FDIS 23008-2: High Efficiency Video Coding (HEVC), JCT-VC Std., Apr. 2013.
Heindel, Wige, Kaup: "SWP for Lossless EL Coding in HEVC"
Chair of Multimedia Communications and Signal Processing
Page 16
Jan. 30, 2014
Coding Efficiency Results
Relative bitrate difference for EL coding with HM-11.0 and SELC compared to the EL of SHM-2.1.
Heindel, Wige, Kaup: "SWP for Lossless EL Coding in HEVC"
Chair of Multimedia Communications and Signal Processing
Page 17
Jan. 30, 2014
Encoding Times
Relative runtime increase for EL encoding compared to BL encoding only.
Heindel, Wige, Kaup: "SWP for Lossless EL Coding in HEVC"
Chair of Multimedia Communications and Signal Processing
Page 18
Jan. 30, 2014
Decoding Times
Relative runtime increase for EL decoding compared to BL decoding only.
Heindel, Wige, Kaup: "SWP for Lossless EL Coding in HEVC"
Chair of Multimedia Communications and Signal Processing
Page 19
Jan. 30, 2014