speed up in feedback channel for a ldpca base distributed video coding system on mobile device
DESCRIPTION
Speed up in feedback channel for a LDPCA base distributed video coding system on mobile device. 在手機裝置上對低 密度奇偶校驗碼為基礎之分散式 編碼中的回饋通道加速. Chen,chun -yuan 陳群元 Advisor: Prof. Wu, Ja -Ling 吳家麟 教授 2012/5/18. outline. Motivation and introduction Mobile video trans Traditional video codec - PowerPoint PPT PresentationTRANSCRIPT
Speed up in feedback channel for a LDPCA base distributed video coding
system on mobile device在手機裝置上對低密度奇偶校驗碼為基礎之分散式編碼中的回饋通道加速
Chen,chun-yuan 陳群元Advisor: Prof. Wu, Ja-Ling 吳家麟 教授
2012/5/18
outline Motivation and introduction Mobile video trans
Traditional video codec Distributed video codec
DVC architecture Channel coding Ldpca
The video communication system on Mobile device Speed up DVC decoding time in feedback channel Experiment Result Conclusion
Digital Video Service
Video compression is an essential component of broadcast and entertainment media
Multimedia video everywhere!
mobile video conference
mobile cameras phone
Video surveillance
Wireless sensor network
outline Motivation and introduction Mobile video trans
Traditional video codec Distributed video codec
DVC architecture Channel coding Ldpca
The video communication system on Mobile device Speed up DVC decoding time in feedback channel Experiment Result Conclusion
Conventional Video Codec
MPEG-2, H.264, HEVC(H.265)ENCODER DECODERLightweig
htHeavyweight
Application of DVC Video conferencing with mobile
devices
DVC to H.264 Transcoder
CloudComputational Resource
DVC encoder(Low Complexity)
H.264 decoder(Low Complexity)
DVC encoded bitstream
H.264 encoded bitstream
outline Motivation and introduction Mobile video trans
Traditional video codec Distributed video codec
DVC architecture Channel coding Ldpca
The video communication system on Mobile device Speed up DVC decoding time in feedback channel Experiment Result Conclusion
Distributed Video Coding
RY =H (Y)
Source encoder
Correlation is exploited by
motion estimation
RY =H (X|Y)Source encoder
H (X|Y)H (Y)-------Distributed-------
Slepian-Wolf Theorem(1973) Wyner-Ziv Theorem(1976)
Channel encoder
Source decoder
Side Information creation
Channel decoder
X
Y
Virtual channelVirtual channel
H (X, Y)
Distributed Video Coding
D. Varodayan, A. Aaron, and B. Girod, “Rate-Adaptive Codes for Distributed Source Coding,”EURASIP Signal Processing Journal, Special Issue on Distributed Source Coding,,November 2006.
Channel Encoder
Channel Decoder
LDPCEncoder
LDPCDecoder
Distributed Video Coding
D. Varodayan, A. Aaron, and B. Girod, “Rate-Adaptive Codes for Distributed Source Coding,”EURASIP Signal Processing Journal, Special Issue on Distributed Source Coding,,November 2006.
Channel Encoder
Channel Decoder
LDPCEncoder
LDPCDecoder
Key frame
Key frame
WZ frame
WZ frame
WZ frame
GOP size 4
Key frame
Key frame
WZ frame
GOP size 2
Quantization
DCT coefficients bands
DCT coefficient band b1 : { S11, S21, S31, …SN1 }DCT coefficient band b2 : { S12, S22, S32, …SN2 }
DCT coefficient band b16 : { S116, S216, S316, …SN16 }
…
DC band
AC bands
Block1
S11 S12 S16 S17
S13 S15 S18 S113
S14 S19 S112 S114
S110 S111 S115 S116
Block2
S21 S22 S26 S27
S23 S25 S28 S213
S24 S29 S212 S214
S210 S211 S215 S216
Block3
S31 S32 S36 S37
S33 S35 S38 S313
S34 S39 S312 S314
S310 S311 S315 S316
…
Distributed Video Coding
D. Varodayan, A. Aaron, and B. Girod, “Rate-Adaptive Codes for Distributed Source Coding,”EURASIP Signal Processing Journal, Special Issue on Distributed Source Coding,,November 2006.
Channel Encoder
Channel Decoder
LDPCEncoder
LDPCDecoder
Quantization
Q8
2^7
2^6
2^6
2^5
2^5
2^5
2^4
2^4
2^4
2^4
2^3
2^3
2^3
2^2
2^2
63bits for one block
Q8
2^52^4
2^4
2^3
2^2
2^32^2
2^32^22^2
2^22^2
2^2
37bits for one block
Bit plane Extraction
00100 00001
00000 11110
Bit planes of DC band:
Bit plane 1:
Bit plane 2:
Bit plane 3:
Bit plane 4:
Bit plane 5:
Independently Channel Encode (LDPCA)
4 6
7
0 6
3
1 7
7
30 1
5
For each DCT coefficient band…
MSB
LSBZig-zag order
outline Motivation and introduction Mobile video trans
Traditional video codec Distributed video codec
DVC architecture Channel coding Ldpca
The video communication system on Mobile device Speed up DVC decoding time in feedback channel Experiment Result Conclusion
LDPC decodingSum-Product Algorithm
(Message Passing)
Side Information(real number)+ 0 - 1
4 6 7
甲 乙 丙
3 521
decode outputhard decision
a b c d e f g
a25 b25 c25 d25 e25 f25 g25
Vertical processing
Horizontalprocessing
a1 b1 c1 d1 e1 f1 g1
0 1 1
From DVC encoder(syndrome bits)
Kschischang, F.R., Frey, B.J., and Loeliger, H.-A. 2001. Factor graphs and the sum-product algorithm. IEEE Trans. Inform. Theory
LDPC Accumulate (LDPCA) codes
Rate adaptivity
D. Varodayan et al., "Rate-adaptive codes for distributed source coding," EURASIP Signal Processing Journal, Special Section on Distributed Source Coding, 2006
65
LD
PC
cod
es
3
outline Motivation and introduction Mobile video trans
Traditional video codec Distributed video codec
DVC architecture Channel coding Ldpca
The video communication system on Mobile device Speed up DVC decoding time in feedback channel Experiment Result Conclusion
Distributed Video Coding
D. Varodayan, A. Aaron, and B. Girod, “Rate-Adaptive Codes for Distributed Source Coding,”EURASIP Signal Processing Journal, Special Issue on Distributed Source Coding,,November 2006.
Channel Encoder
Channel Decoder
LDPCEncoder
LDPCDecoder
Distributed Video Coding
D. Varodayan, A. Aaron, and B. Girod, “Rate-Adaptive Codes for Distributed Source Coding,”EURASIP Signal Processing Journal, Special Issue on Distributed Source Coding,,November 2006.
Channel Decoder
LDPCDecoder
Key frame
request
WZ Bit stream1’th bitstream&CRC
63’th bitstream&CRC
outline Motivation and introduction Mobile video trans
Traditional video codec Distributed video codec
DVC architecture Channel coding Ldpca
The video communication system on Mobile device Speed up DVC decoding time in feedback channel Experiment Result Conclusion
Amdahl's law
Maximum speedup can be reached by improving the most critical part of the system
Feedback channel in the DVC coding.
time ratio
SI generationLDPCA DecFeedback channeletc.
29%
66.6%
Amdahl's law
Maximum speedup can be reached by improving the most critical part of the system
Feedback channel in the DVC coding.
time ratio(baseline)
SI generationLDPCA Dec+Feedback channeletc.95.6%
Amdahl's law
Maximum speedup can be reached by improving the most critical part of the system
Feedback channel in the DVC coding. estimate syndromes size per bitplanetime ratio(estimate syndromes size
per bitplane)
SI generation LDPCA Dec+Feedback channel
etc.92.8%
1.7 X
Amdahl's law
Maximum speedup can be reached by improving the most critical part of the system
Feedback channel in the DVC coding. estimate syndromes size per WZ frametime ratio(estimate syndromes size
per WZ frame)
SI generation LDPCA Dec+Feedback channel
etc.82.8%
4.6 X
Methods to speed up at feedback channel
So we propose two methods to decrease time consuming at feedback channel. Estimate the syndromes size per WZ
frame Estimate the syndromes size per bitplane
Number of requests per bitplane per band
1234567123456123456123451234512345123412341234123412312312312120
200
400
600
800
1000 bit rate
bit rate
DC AC1 AC2 AC3 AC4 AC5 AC6 AC7 AC8AC9AC10AC11
AC12AC13
AC14
Bitplane Number
Number of requests per bitplane per band
1234567123456123456123451234512345123412341234123412312312312120
200
400
600
800
1000 bit rate
bit rate
DC AC1 AC2 AC3 AC4 AC5 AC6 AC7 AC8AC9AC10AC11
AC12AC13
AC14
Bitplane Number
Estimate by previous bands
1.formula 2.time ratio
Number of requests per bitplane per band
1234567123456123456123451234512345123412341234123412312312312120
200
400
600
800
1000 bit rate
bit rate
DC AC1 AC2 AC3 AC4 AC5 AC6 AC7 AC8AC9AC10AC11
AC12AC13
AC14
Bitplane Number
Number of requests per bitplane per band
1234567123456123456123451234512345123412341234123412312312312120
200
400
600
800
1000 bit rate
bit rate
DC AC1 AC2 AC3 AC4 AC5 AC6 AC7 AC8AC9AC10AC11
AC12AC13
AC14
Bitplane Number
Number of requests per bitplane per band
1234567123456123456123451234512345123412341234123412312312312120
200
400
600
800
1000
bit rateestimated bit rate
DC AC1 AC2 AC3 AC4 AC5 AC6 AC7 AC8AC9AC10AC11
AC12AC13
AC14
Bitplane Number
formula
(ESacn bt)WZn= ( Sdc bt-1 )WZn
WZn means the n’th WZ frame. acn means the n’th AC band. bt means the bitplane index of this AC band. Which ESacn
bt mean the estimated syndrome size for the bt’th bitplane in n’th AC band.
dc means the DC band. Sdc bt-1 means the syndromes bits for bt-1’th
bitplane in DC band.
Amdahl's law
Maximum speedup can be reached by improving the most critical part of the system
Feedback channel in the DVC coding.
time ratio(baseline)
SI generationLDPCA Dec+Feedback channeletc.95.6%
406.39 S
Amdahl's law
Maximum speedup can be reached by improving the most critical part of the system
Feedback channel in the DVC coding.
time ratio(estimate syndromes size per bitplane)
SI generation LDPCA Dec+Feedback channel
etc.92.8%
1.7 X
237.87 S
Estimate by reference frame
1.formula 2.time ratio
formula
(ESbt)WZn=(Sbt)WZ(n-GOPsize)
WZn means the WZ frame’s index n-GOPsize means the correspond WZ
frame in the previous GOP. bt is the bitplane index. (ESbt)WZn means the estimated
syndromes bit for the bt’th bitplane of WZn frame.
(Sbt)WZ(n-GOPsize) means the syndromes
Amdahl's law
Maximum speedup can be reached by improving the most critical part of the system
Feedback channel in the DVC coding.
time ratio(baseline)
SI generationLDPCA Dec+Feedback channeletc.95.6%
406.39 S
Amdahl's law
Maximum speedup can be reached by improving the most critical part of the system
Feedback channel in the DVC coding.
time ratio(estimate syndromes size per WZ frame)
SI generation LDPCA Dec+Feedback channel
etc.82.8%
4.6 X
87.99 S
outline Motivation and introduction Mobile video trans
Traditional video codec Distributed video codec
DVC architecture Channel coding Ldpca
The video communication system on Mobile device Speed up DVC decoding time in feedback channel Experiment Result Conclusion
Test condition
12 CPU, 24 processor Intel(R) Xeon(R) CPU X5650 @ 2.67GHz
GPU: Tesla M2050 Mobile device: HTC sensation
DVC encoder on mobile
14 s for foreman sequence, GOP8 , intra mode on.
4 s for foreman sequence, GOP8 , intra mode off.
Test sequences :
QCIF, 15Hz, all frames GOP Size 2, 4 and 8 Only luminance component is used
Test materials
Soccer Foreman Coastguard Hall Monitor
Motion: High Low
Speed up
Bitrate
RD curve
Foreman with LRSS,GOP8
RD curve
Hall monitor with LRSS,GOP8
outline Motivation and introduction Mobile video trans
Traditional video codec Distributed video codec
DVC architecture Channel coding Ldpca
The video communication system on Mobile device Speed up DVC decoding time in feedback channel Experiment Result Conclusion
conclusion
A practical DVC coding system is built on mobile devices. DVC encoder and H.264 decoder keep
the mobiles light weight. We propose two method to reduce
transmission time spent in feedback channel. Estimate the syndromes size per WZ
frame Estimate the syndromes size per bitplane
Thank you