Project Proposal

Title : Optimal Mode Selection For Robust Video Transmission

Name: Sangoh Jeong
E-mail: sojeong@stanford.edu


  The conundrum in today's video communication system is to achieve
the robustness of compressed video data in error-prone environment. And 
a variety of techniques are being tackled not only in academia but also
in industry. However, due to the innate vulnerability of compressed video
data to the transmission errors, it doesn't seem that the reliable video
communication in error-prone environment is so near-at-hand. And this
proposal is another trial to overcome that barrier by way of a new approach.

  The goal of this project is to propose a new coding mode(in-between mode)
in hybrid encoder and to use a new control scheme using that mode. Theoreti-
cally, the first is based on leaky prediction in hybrid encoder and the
latter is founded on the well-known rate-distortion theory. The proposed
coding scheme considers the trade-off between coding efficiency and error
resiliency. It is also expected to show better characteristics than the 
traditional encoding control operation using only intra and inter coding
modes in error-prone environment. And the whole encoding operation depending
on the new coding control scheme requires only moderate complexity if it is
compared with the entire encoding process which uses two-mode(intra,inter)
-only control scheme.

  In the experiment, much of the work will be focused on the error-robustness
of the new encoding scheme. But, for the purpose of comparison, the experiment
under the two-mode-only control scheme will precede in the first hand. H.263+
(TMN-10) codec software will be used to verify the proposed coding scheme 
since it has both the two-mode-only coding control and rate-distortion based
coding control in its encoder structure. Two kinds of errors are inserted to
the encoded bitstream as transmission errors, one of which is the burst error
and the other is the random bit error. In order to simulate burst errors, it 
is assumed that there will be loss of 1 GOB(Group Of Block) in every two frame.
And, in case of random bit errors, 5% of bits in the bitstream of a frame will
be randomized. The same experiment will be repeated for several known QCIF
video sequences. The final PSNR behaviors will be compared between both of
coding schemes mentioned earlier.



Workplan:

   Feb.2 - Feb.20  : Study rate-distortion theory
		     Test the codec software
		     Finish the experiment using two-mode-only encoding scheme

   Feb.12 - Feb.25 : Design the proposed encoding control scheme 
		     Document intermediate results

   Feb.18 - Mar.10 : Finish the experiment using the proposed encoding scheme

   Mar.5 - Mar.16  : Make the final report
		     Prepare for the presentation



References:

[1] B. Girod and N. Faber, "Wireless Video", in A. Reibman, M.-T. Sun (eds.), 
    Compressed Video over Networks, Marcel Dekker, 2000.
[2] G. J. Sullivan and T. Wiegand, "Rate-Distortion Optimization for Video 
    Compression" IEEE Signal Processing Magazine, November 1998.
[3] R. Zhang, S. L. Regunathan and K. Rose, "Video Coding with Optimal 
    Inter/Intra-Mode Switching for Packet Loss Resilience", IEEE Journal of 
    Selected Areas in Communications, Vol. 18, NO. 6, June 2000
[4] G. Cote, S. Shirani, and Faouzi Kossentini, "Optimal Mode Selection and 
    Synchornization for Robust Video Communications over Error-Prone Networks",
    IEEE Journal of Selected Areas in Communications, Vol. 18, NO. 6, June 2000