1.
Wyner-Ziv Video Coding With Improved
Motion Field Using Bicubic Interpolation
I Made Oka Widyantara
Telecomunication System Lab.
Department of Electrical Engineering
Udayana University

2.
Outline
 Introduction
 Proposed Wyner-Ziv Video codec
 Bicubic Interpolation of motion field
 Experiments and Results
 Analysis of RD performance
 Analysis of decosing complexity
 Conclusions

3.
Introduction
 WZVC is the recent video coding paradigm based on
the Slepian-Wolf, and Wyner-Ziv theorems which
exploits the source temporal correlation at the
decoder and not at the encoder.
 The main problem in designing WZVC codec is a
method of generating the SI at the decoder

4.
Introduction
 Motion vector learning based on Expectation Maximization
(EM) algorithm is a method of generating SI for WZVC
iteratively (Varodayan et.al., 2008).
Rate control
X S θ
LDPC
Encoder
LDPC Dedoder
(M-Step)
ψ
Probability
Model
Reconstruction
^
X
Block-based
motion estimator
(E-step)
Ŷ
Motion field
interpolation
P{Mi,j} P{Mu,v}
 Motion field interpolation : refines the block based motion
field P{Mu,v}to pixel precision P{Mi,j}, to improve the
accuracy of the soft SI.
 We proposed bicubic interpolation techniques, and make
performance analysis WZVC codec, when compared with
the implementation of bilinear interpolation technique
(WZVC existing)

5.
Proposed WZVC codec
 Initialize block-based motion field : 
 


if M
, (0,0)
3
3
P M if M
u v
u v
   
80 ,
  


otherwise
u v
t
app
,
(0, ), ( ,0)
2
1
80
1
4
,
2
4
,
( )

6.
Bicubic Interpolation of motion
field
 Uses sixteen probability distribution Papp{Mu,V(xS,yS)} that close to
(xS,yS) position in block based motion field.
 First, for each ysk, the algorithm determines four polynomial cubic
F0(x), F1(x), F2(x), and F3(x) using:
Fk(x) = akx3 + bkx2 + ckx + dk, ; 0 ≤ k ≤ 3
such that : Fk(xS0) = Papp{Mu,v(xSo,ySk)}, Fk(xS1) = Papp{Mu,v(xS1,ySk)},
Fk(xS2) = Papp{Mu,v(xS2,ySk)}, Fk(xS3) = Papp{Mu,v(xS3,ySk)}
 Then, the algorithm determines a cubic polynomial Fy(y) such that:
Fy(yS0) = F0(xS), Fy(yS1) = F1(xS), Fy(yS2) = F2(xS), Fy(yS3) =
F3(xS)
 Finally, the value of probability distribution Papp{Mi,j(xD,yD)} is set on
Fy(yS).
Probability
F0(x)
F1(x)
F2(x) F3(x)
x
y
Papp{Mi,j(xS,yS)}
xS
xS3
yS0 yS1 yS yS2 yS3
xS0
xS1
xS2

7.
Analysis of RD performance
RD curves for GOP sizes 2, Foreman (left), Carphone (right)
 WZVC codec with Bicubic interpolation produces an almost
identical RD performance with existing WZVC codec with
Bilinear interpolation.
 At fixed rate, both interpolation techniques produce the same
PSNR gain and constant throughout scaling factor quantization,
Qf = 0.5, 1, 2 and 4.
 Identical RD performance produced by both methods showed
that Bicubic interpolation is able to produce a linear convex
combination of probability on the interval [0,1].

8.
Analysis of decoding complexity
Foreman
Carphon
e
 The decoding complexity is evaluated by measuring the average
decoding time per quadrant of EM iteration time needed by
decoder to fulfill the conditions of syndrome.
 In general, for both video sequences used, the implementation of
Bicubic interpolation reduces the complexity of the decoder
WZVC
 The most decrease in complexity occurred in scaling factor Qf =
0.5 up to 9.49% for Foreman, and up to 7.33% for Carphone.
This indicates that the codec WZVC with Bicubic interpolation is
suitable to encode video sequences with high and complex
motion content.

9.
Conclusions
 The new WZVC codec improves motion field
probability distribution into pixel precision, using
Bicubic interpolation technique.
 Experimental results showed that implementation
of bicubic interpolation technique reduces
decoder complexity significantly with RD quality
almost equal to previous learning based WZVC
codec that use Bilinear interpolation.

10.
Thank You