Sale!

EE6427 Assignment

$30.00

Category:

Description

5/5 - (2 votes)

EE6427 Assignment

I will let you know how to submit the assignment at a later date.
(NO late submission is allowed)
ALL works submitted must be based on your individual effort, please don’t copy other
people’s results.
(1) Calculate two-dimensional transform of figure 1 by using row-column decomposition
method with the basis function in figure 2, please show all the intermediate steps to
obtain the final result.






10 10 10 10 10 10 10 10
10 10 10 10 10 10 10 10
20 20 20 20 20 20 20 20
20 20 20 20 20 20 20 20
40 40 40 40 40 40 40 40
40 40 40 40 40 40 40 40
10 10 10 10 10 10 10 10
10 10 10 10 10 10 10 10⎠





Figure 1
Figure 2
if the quantization matrix on page 6 of the lecture note “JPEG” is used, calculate the
quantization output. What is the one-dimensional output after zig-zag scanning?
(20 marks)
(2) Define a character string using YOUR FULL NAME as appearing in Matriculation Card
(all in capital letters and remove all spaces in your name) and follow by
“VIDEOSIGNALP”.
Use arithmetic coding method in the lecture note “compression fundamental” to
encode the first 8 letters of the character string. Show the steps of the divisions of the
interval during arithmetic encoding the character string, and show the codeword
produced by the encoding procedure.
For example, “CHA TAI” should encode “CHATAIVI”. You need to use YOUR FULL NAME
as appearing in Matriculation Card (all in capital letters and remove all spaces in your
name). And assign the interval (range) in alphabetical order. E.g. “A” should be assigned
as the first letter in the interval. (please refer to the example in the lecture note). (Zero
mark will be given if you don’t follow the rule.)
(20 marks)
(3) Define a character string using YOUR FULL NAME as appearing in Matriculation Card
(all in capital letters and remove all spaces in your name) and follow by
“VIDEOSIGNALP”. Please remove all spaces in the string and let “A”=1, “B”=2, …,
“Y”=25, “Z”=26 as an input to fill up the following 4×4 matrix. (Zero mark will be given if
you don’t follow the rule.)
For example : If YOUR FULL NAME as appearing in Matriculation Card, e.g. “CHA TAI”,
the 16 letters are “CHATAIVIDEOSIGNA” the corresponding numbers for the first 16
letters are “ 3 8 1 20 1 9 22 9 4 5 15 19 9 7 14 1”
3 8 1 20
1 9 22 9
4 5 15 19
9 7 14 1
Let the 4×4 matrix (obtained from above) be a two-level discrete wavelet transform
decomposition result. Applying the EZW coding scheme to the wavelet coefficients and
show the encoding result. Note that four symbols in dominant pass for EZW are T
(zerotree root), Z (isolated zero), P (positive) and N (negative) respectively.
(20 marks)
(4) Plot the rate distoration curve by changing the parameters at an H.263 encoder. To
do this question, you need to explore the option of the tmn software by yourself.
• Use the “football_cif.yuv” (can be obtained from ntulearn site) sequence of first
150 frames.
• Use tmn.exe (Unzip from h263.zip. It is a DOS program which requires to execute
on command prompt) to generate the result.
• Perform experiments with different quantization parameters, try at least 20
different QPs to obtain meaning results.
Discuss your results based on the reconstructed video quality and the MSE obtained.
You may need to write a program to calculate the overall PSNR and MSE, where i x are
the original pixels and ˆi x are the reconstructed pixels obtained from tmndec. Please
show all the steps to obtain the results.
Plot the PSNR-Y against various bitrate as shown in figure 3.
Figure 3
Plot the MSE-Y against various bitrate as shown in figure 4.
Figure 4
Fix the bitrate to different values (at least 5 different bitrates), plot the MSE-Y against
frame number as shown in the following figure 5.
Figure 5
Please comment on your results.
(40 marks)
Bitrate
PSNR
Frame no
MSE
Bitrate
MSE