ArticlePDF Available

Figures

Content may be subject to copyright.
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 2 Issue: 8 2277 2284
_______________________________________________________________________________________________
2277
IJRITCC | August 2014, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
Wavelet Based Performance Analysis of Image Compression
Md. Taslim Arefin1, Md. Zahirul Islam1, Md. Asaduzzaman Khan2, Md. Mahfujur Rahman1 , A.S.M Shaem1
1Dept. of Electronics and Telecommunication Engineering
Faculty of Engineering
Daffodil International University
Dhaka, Bangladesh
E-mail: arefin@daffodilvarsity.edu.bd,
zahirete@daffodilvarsity.edu.bd
mahfuj.ete@daffodilvarsity.edu.bd
2Dept of Computer Science and Engineering
Leading University
Sylhet, Bangladesh
E-mail: nepon_1979@yahoo.com
AbstractIn this paper, our aim is to compare for the different wavelet-based image compression techniques. The effects of different wavelet
functions filter orders, number of decompositions, image contents and compression ratios were examined. The results of the above techniques
WDR, ASWDR, STW, SPIHT, EZW etc., were compared by using the parameters such as PSNR, MSE BPP values from the reconstructed
image. These techniques are successfully tested by four different images.
Keywords- Image compression, WDR, STW, PNSR
__________________________________________________*****_________________________________________________
I. INTRODUCTION
Present day large amounts of images are stored, processed
and transmitted and hence there is a great need for the
compression of an image to save memory, transmission
bandwidth etc. For many applications, simply reducing the file
size or simple compression is not sufficient some additional
scalable and embedded properties are also required. Discrete
Wavelet Transform (DWT) provides a multi resolution image
representation and has become one of the most important tools
in image analysis and coding over the last two decades.[1]
Wavelet transforms have been widely studied over the last
decade. At the present state technology, the only solution is to
compress multimedia data before its storage and transmission,
and decompress it at the receiver for playback. For example for
a compression Ratio of 32:1 ,the space, bandwidth and the
transmission time requirements can be reduced by a factor of
32,with acceptable quality. The fundamental goal of image
compression is to reduce the bit rate for transmission or storage
while maintaining an acceptable fidelity or image. One of the
most successful applications of wavelet methods is transform-
based image compression (also called coding).Wavelet-based
coding provides substantial improvements in picture quality at
higher compression ratios. This paper presents the Analysis of
different wavelet based image compression Techniques. The
existing Techniques, WDR, ASWDR, STW, SPHIT, EZW
have been introduce and evaluated based on the parameters like
PSNR, MSE, BPP. Acceptable image quality has been
extracted in terms of the performance parameter and coding
technique [3]. The result is extracted from the experiment
empirically and shows that the EZW and STW technique
performs better than WDR and other method in terms of the
parameters. The analysis has been tested and verified Using
MATLAB.
II. WORKING METHODOLOGY
Wherever Times is specified, Times Roman or Times New
Roman may be used. If neither is available on your word
processor, please use the font closest in appearance to Times.
As a mathematical tool, wavelets can be used to extract
information from many different kinds of data, including but
certainly not limited to audio signals and images. Sets of
wavelets are generally needed to analyze data fully. A set of
"complementary" wavelets will deconstruct data without gaps
or overlap so that the deconstruction process is mathematically
reversible. Thus, sets of complementary wavelets are useful in
wavelet based compression/decompression algorithms where it
is desirable to recover the original information with minimal
loss. There are many compression methods in wavelet section
like:
EZW (Embedded Zero tree Wavelet)
SPIHT (Set Partitioning in Hierarchical Trees)
STW (Spatial-orientation Tree Wavelet)
WDR (Wavelet Difference Reduction)
ASWDR (Adaptively Scanned Wavelet Difference
Reduction)
We have use those methods in this work and also checked
performance analysis in different situation. The basic scheme
for compressing images is shown in Figure 1 below.
Compression consists of two steps to generate a compressed
bit stream.
The rest step is a wavelet transform of the image and the
second step is the compressed encoding of the image’s wavelet
transform. Decompression simply consists of reversing these
two steps, decoding the compressed bit stream to produce an
(approximate) image transform. In the block diagram the total
procedure is shown by flow chart in figure 1 and then
described.
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 2 Issue: 8 2277 2284
_______________________________________________________________________________________________
2278
IJRITCC | August 2014, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
Fig 1: Basic scheme for compressing images
At first some image has been taken from the camera.
Then it’s sent through the MATLAB basement and then
resizes it to ―512*512*3 ―format because we know that
for true compression, it is necessary to keep the size of
rows and columns in the power of 2.
Then take the Haar wavelet for compression then apply
different types of method like as EZW, SPHIT, WDR,
ASWDR, STW etc.
Then It has been compared between different method
and select the best method for compression here 4 photos
are taken and then analyst it.
Figure 2a: Screen shot of the wavelet menu
Figure 2b: MATLAB 2-D image compression interface
Wavelet-based coding provides substantial improvements
in picture quality at higher compression ratios. Over the past
few years, a variety of powerful and sophisticated wavelet-
based schemes for image compression, as discussed later, have
been developed and implemented. Because of the many
advantages, wavelet-based compression algorithms are the
suitable candidates for the new JPEG-2000 standard. This is
lossy compression. In many cases, it is not necessary or even
desirable that there be error-free reproduction of the original
image. Lossy compression is also acceptable in fast
transmission of still images over the Internet. Over the past few
years, a variety of novel and sophisticated wavelet-based image
coding schemes have been developed. These include
Embedded Zero tree Wavelet (EZW), Set-Partitioning in
Hierarchical Trees (SPIHT), Wavelet Difference Reduction
(WDR), Adaptively Scanned Wavelet Difference Reduction
(ASWDR), and STW. This list is by no means exhaustive and
many more such innovative techniques are being developed. A
few of these algorithms are briefly discussed here.
III. PERFORMANCE ANALYSIS
In this section the overall performance analysis will be
discussed using different wavelet method like as EZW,
SPIHT, WDR, ASWDR and STW etc. Before starting the
work, our purpose in discussing the baseline compression
algorithm was to introduce some basic concepts, such as scan
order, effects of different wavelet functions filter orders,
number of decompositions, image contents and compression
ratios, P.S.N.R, B.P.P were examined, which are needed for
my examination of the algorithms to follow.
4.1 Embedded Zero tree Wavelet (EZW): The EZW
algorithm was one of the first algorithms to show the full
power of wavelet-based image compression. It was introduced
in the groundbreaking paper of Shapiro. An EZW encoder is
an encoder specially designed to use with wavelet transforms.
The EZW encoder is based on progressive encoding to
compress an image into a bit stream with increasing accuracy
[5].
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 2 Issue: 8 2277 2284
_______________________________________________________________________________________________
2279
IJRITCC | August 2014, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
TABLE-1: Compression ratio Bit per pixel and PSNR Result for 512*512*3
image.
Size
Level
CR
PSNR
BPP
512*512
2
81.58
58.69
6.528
512*512
2
89.04
59.76
7.1232
512*512
2
58.07
51.95
4.6458
512*512
2
74.17
57.85
5.93
Figure: 3a Compress screenshot of four images using in EZW method
4.2 Set Partitioning in Hierarchical Trees (SPIHT)
SPIHT is a wavelet-based image compression coder. It first
converts the image into its wavelet transform and then
transmits information about the wavelet coefficients. The
decoder uses the received signal to reconstruct the wavelet and
performs an inverse transform to recover the image. We
selected SPIHT because SPIHT and its predecessor, the
embedded zero tree wavelet coder, were significant
breakthroughs in still image compression in that they offered
significantly improved quality over vector quantization,
JPEG, and wavelets combined with quantization, while not
requiring training and producing an embedded bit stream [4] .
TABLE-2: Compression ratio Bit per pixel and PSNR Result for (512*512*3)
image.
Level
CR
PSNR
BPP
2
46.71
38.87
3.737
2
47.53
39.32
3.8028
512*512
2
39.16
39.9
3.13
512*512
2
35.73
40.74
2.8585
Figure: 3b Compress screenshot of four images using in SPIHT method
Wavelet Difference Reduction (WDR): One of the defects of
SPIHT is that it only implicitly locates the position of
significant coefficients. This makes it difficult to perform
operations which depend on the position of significant
transform values, such as region selection on compressed data.
Region selection, also known as region of interest (ROI),
means a portion of a compressed image that requires increased
resolution [2].
TABLE-3: Compression ratio Bit per pixel and PSNR Result for (512*512*3)
image.
Size
Level
CR
PSNR
BPP
512*512
2
80.02
40.16
6.4012
512*512
2
89.14
41.33
7.131
512*512
2
67.23
41.61
5.37
512*512
2
62.00
42.84
4.95
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 2 Issue: 8 2277 2284
_______________________________________________________________________________________________
2280
IJRITCC | August 2014, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
Figure: 3c Compress screenshot of four images using in WDR method
Adaptively Scanned Wavelet Difference Reduction
(ASWDR)
The ASWDR algorithm aims to improve the subjective
perceptual qualities of compressed images and improve this
result of objective distortion measures. We shall treat two
distortion measures, PSNR and edge correlation, which we
shall define in the section or experimental results. PSNR is a
commonly used measure of error, while edge correlation is a
measure that we have found useful in quantifying the
preservation of edge details in compressed images, and seems
to correspond well to subjective impressions of the perceptual
quality of the compressed images. [2]
TABLE-4: Compression ratio Bit per pixel and PSNR Result for (512*512*3)
image.
Size
Level
CR
PSNR
BPP
512*512
2
76.22
40.16
6.0975
512*512
2
84.64
41.33
6.7708
512*512
2
62.88
41.61
5.0306
512*512
2
59.27
42.84
4.7419
Figure: 3d Compress screenshot of four images using in ASWDR method
4.5 Spatial-orientation Tree Wavelet (STW)
STW is essentially for the SPIHT algorithm. The only
difference is that SPIHT is slightly more careful in its
organization of coding output. Second, we describe the SPIHT
algorithm. It is easier to explain SPIHT using the concepts
underlying STW. Third, we see how well SPIHT compresses
images. The only difference between STW and EZW is that
STW uses a different approach to encoding the zero tree
information. STW uses a state transition model. From one
threshold to the next, the locations of transform values
undergo state transitions. This model allows STW to reduce
the number of bits needed for encoding. Instead of code for the
symbols R and I output by EZW to mark locations, the STW
algorithm uses states IR, IV , SR, and SV and outputs code for
state-transitions such as IR → IV , SR → SV , etc.
TABLE-5: Compression ratio Bit per pixel and PSNR Result for (512*512*3)
image.
Size
Level
CR
PSNR
BPP
512*512
2
54.37
47.52
4.3497
512*512
2
58.55
45.91
4.68
512*512
2
41.87
46.53
3.34
512*512
2
37.74
45.45
3.019
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 2 Issue: 8 2277 2284
_______________________________________________________________________________________________
2281
IJRITCC | August 2014, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
Figure: 3e Compress screenshot of four images using in STW method
IV. SIMULATION AND RESULTS
In this paper it have been said that the overall performance
of the four images are shown in the below table. Here we have
been said that, in case of EZW method MSE is 0.0868 &
PSNR is 58 & BPP is 6.5267. For SPIHT method, MSE is
7.04 & PSNR is 39.50 & BPP is 3.379. For STW method,
MSE is 1.6& PSNR is 45.91 & BPP is 3.84. For WDR
method, MSE is 4.72& PSNR is 41.29& BPP is5.97. For
ASWDR method, MSE is 4.752 & PSNR is 41.49 & BPP
is5.660.Among all these methods, EZW are best performed
though STW is averagely good as compared to the other
method.
TABLE-6: Total Analytical Result for (512*512*3) image.
EZW
(a,b,c,d)
SPIHT
(a,b,c,d
)
STW
(a,b,c,d
)
WDR
(a,b,c,d
)
ASWD
R
(a,b,c,d)
M.S.E
0.08782
8.431
1.152
6.27
6.27
0.6868
7.597
1.668
4.785
4.875
0.415
6.651
1.445
4.492
4.492
0.1068
5.485
1.853
3.378
3.378
P.S.N.
R
58.69
38.87
47.52
41.33
40.16
59.76
39.32
45.91
39.38
41.33
51.95
39.1
46.53
41.61
41.61
57.85
40.74
45.45
42.84
42.84
B.P.P
6.5267
3.737
4.3497
6.4012
6.0975
7.1232
3.8028
4.684
7.131
6.7708
4.6458
3.1327
3.3497
5.3781
5.0306
5.9339
2.8585
3.0191
4.9598
4.7419
Figure: 4 Graph 1(For total analytical result)
The graph in figure 4 represents that the overall performance
for different method. Here it has been seen that among the five
methods, EZW perform better than other method here MSE of
EZW method is lower and peak signal to noise ratio is higher
than the other method. The below table state that the average
performance for five method here different types of parameter
like as compression ratio, Peak signal to noise ratio and Bit per
pixel have been discussed.
TABLE-7: Average Result for 4 (512*512*3) image
Figure: 5a Graph 2 (For BPP vs.MSE)
In figure 5a the graph represents that the average compares
between Bit per pixel and Mean square error for different
method .here it have been seen that for EZW method the mean
square error is lesser then the another method where the Bit
per pixel is medium.
Figure: 5b Graph 3 (For BPP vs. PSNR)
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 2 Issue: 8 2277 2284
_______________________________________________________________________________________________
2282
IJRITCC | August 2014, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5
35
40
45
50
55
60 Bit per pixel vs peak signal to noise ratio for ALL method
Bit per pixel
peak signal to noise ratio
EZW
SPIHT
STW
WDR
ASWDR
In the figure 5b two types of parameter have been discussed
and compare between them here it has been seen that the peak
signal to noise ratio is higher than the others method where the
Bit per pixel is lower
Figure: 5c Graph 4 (Overall performance for CR, PSNR, BPP,
MSE)
In the graph in figure 5c represents that the overall
performance for CR, PSNR, BPP, MSE. And it have been seen
that for EZW method BPP is lower where PSNR is higher
than the other method so in case of the total overall
performance analysis it has been said that EZW perform
better than other.
MATLAB Analysis
The figure 6a shows B.P.P VS. MSE The simulation result of
the graph represents that EZW performed best among all other
method .Here B.P.P & M.S.E is less for EZW and all the other
method does not perform as well.
The Figure 6b shows B.P.P VS PSNR The simulation result of
the second graph represents that EZW is perform best as
compared to other method. Here B.P.P is less as well M.S.E. is
less that time P.S.N.R is high and it shows maximum output
performance.
Figure 6a: Bit per Pixel vs. Mean Square error.
Figure :6b: Bit per Pixel vs. Peak Signal to Noise Ratio.
Figure 7a: BPP vs. MSE for EZW method
Figure 7b: BPP vs. MSE for SPIHT method
Figure 7c : BPP vs. MSE for STW method
2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5
0
1
2
3
4
5
6
7
8
9Bit per pixel vs mean square error for ALL method
Bit per pixel
mean square error
EZW
SPIHT
STW
WDR
ASWDR
4.5 5 5.5 6 6.5 7 7.5
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7 Bit per pixel vs mean square error for EZW method
Bit per pixel
mean square error
2.8 3 3.2 3.4 3.6 3.8 4
5
5.5
6
6.5
7
7.5
8
8.5 Bit per pixel vs mean square error for SPIHT method
Bit per pixel
mean square error
3 3.2 3.4 3.6 3.8 4 4.2 4.4 4.6 4.8
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9 Bit per pixel vs mean square error for STW method
Bit per pixel
mean square error
4.5 5 5.5 6 6.5 7 7.5
3
3.5
4
4.5
5
5.5
6
6.5 Bit per pixel vs mean square error for WDR method
Bit per pixel
mean square error
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 2 Issue: 8 2277 2284
_______________________________________________________________________________________________
2283
IJRITCC | August 2014, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
4.5 5 5.5 6 6.5 7
3
3.5
4
4.5
5
5.5
6
6.5 Bit per pixel vs mean square error for ASWDR method
Bit per pixel
mean square error
Figure 7d : BPP vs. MSE for WDR method
Figure 7e: BPP vs. MSE for ASWDR method
Figure: 7 a- 7e : Graphical representation of four images in
five methods. Figure: 8: Graphical representation of BPP vs.
MSE for all Methods
In the figure it have been seen that BPP vs. MSE compares
where in 4.5 to 6 BPP the MSE is lower in EZW method but
other method its higher up to 9.5 where EZW method the MSE
is .08 to 1.5.
Figure 8a : BPP vs. PSNR for EZW method
Figure 8b : BPP vs. PSNR for SPIHT method
Figure 8c: BPP vs. PSNR for STW method
Figure 8d: BPP vs. PSNR for WDR method
Figure 8e : BPP vs. PSNR for ASWDR method
Figure: 8a 8e: Graphical representation of BPP vs. PSNR for
all methods
The figure represents that the Bit per Pixel vs. Peak signal to
noise ratio in different method. Here it have been seen that for
EZW method is performed from 4.5 to 6 the PSNR is higher
and it almost 58.44.In case of other method we see for 4.5 to 6
BPP the PSNR not more than 45 .so Here the individual EZW
method performed better.
From the analysis it have been seen that, the various features
of the main coding schemes are summarized. The latest coding
techniques such as EZW perform better than the other method.
4.5 5 5.5 6 6.5 7 7.5
51
52
53
54
55
56
57
58
59
60 Bit per pixel vs peak signal to noise ratio for EZW method
Bit per pixel
peak signal to noise ratio
2.8 3 3.2 3.4 3.6 3.8 4
38.8
39
39.2
39.4
39.6
39.8
40
40.2
40.4
40.6
40.8 Bit per pixel vs peak signal to noise ratio for SPIHT method
Bit per pixel
peak signal to noise ratio
3 3.2 3.4 3.6 3.8 4 4.2 4.4 4.6 4.8
45
45.5
46
46.5
47
47.5
48 Bit per pixel vs peak signal to noise ratio for STW mwthod
Bit per pixel
peak signal to noise ratio
4.5 5 5.5 6 6.5 7 7.5
39
39.5
40
40.5
41
41.5
42
42.5
43 Bit per pixel vs peak signal to noise ratio for WDR Method
Bit per pixel
peak signal to noise ratio
4.5 5 5.5 6 6.5 7
40
40.5
41
41.5
42
42.5
43 Bit per pixel vs peak signal to noise ratio for ASWDR method
Bit per pixel
peak signal to noise ratio
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 2 Issue: 8 2277 2284
_______________________________________________________________________________________________
2284
IJRITCC | August 2014, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
Here we see that from the EZW method we can get the
maximum peak signal to noise ratio and low Bit per pixel.
V. CONCLUSION
In this paper, the results were compared for the different
wavelet-based image compression techniques. The effects of
different wavelet functions filter orders, number of
decompositions, image contents and compression ratios were
examined. The results of the above techniques WDR,
ASWDR, STW, SPIHT, EZW etc., were compared by using
the parameters such as PSNR, MSE BPP values from the
reconstructed image. These techniques are successfully tested
in many images. The experimental results show that the EZW
technique performs better than the WDR & other method in
terms of the performance parameters and coding time with
acceptable image quality. From the experimental results, it is
identified that the PSNR values from the compressed images
by using EZW compression is higher than other compression.
And also it is shown that the MSE values from the
reconstructed images by using EZW compression are lower
than other compression. Finally, it is identified that EZW
compression performs better when compare to WDR,
ASWDR and other compression
REFERENCES
[1] V. R. Algazi, R. R. Estes. ―Analysis based coding of
image transform and sub band coefficients,‖
Applications of Digital Image Processing XVIII,
volume 2564 of SPIE Proceedings, pages 11-21, 1995.
[2] S.P.Raja1, Dr. A. Suruliandi2, Image Compression
using WDR & ASWDR Techniques with different
Wavelet Codecs, Proc. of Int. Conf. on Advances in
Computer Engineering 2011.
[3] Strang, G. and Nguyen, T. Wavelets and Filter Banks,
Wellesley-Cambridge Press, Wellesley, MA, 1996,
http://www-math.mit.edu/~gs/books/wfb.html.
[4] R.Sudhakar, Ms R Karthiga, S.Jayaraman, Image
Compression using Coding of Wavelet Coefficients A
Survey,ICGST-GVIT journal, volume(5), issue(6),june
2005
[5] James S. Walker, A Lossy Image Codec Based on
Adaptively Scanned Wavelet Difference Reduction.
Chapter
Digital image storage and transmission is a challenging task nowadays. As per statistics, an average of 1.8 billion images are transmitted daily. Hence, image compression is inevitable. Here, we discuss a novel approach, which is a type of lossy image compression. But the quality of reconstructed image is higher. This method makes use of reducing the number of bytes by storing the number of ‘1’s in each bitplane of three adjacent pixels and recording the count of number of ‘1’s. This count value is stored, and later Huffman compression is applied. Reconstruction is done by energy distribution method. This gives better results in terms of quality at lower PSNR values compared to JPEG algorithm. Here we have used another metric to measure the quality of the image which is discussed in detail.
Article
Full-text available
This paper describes a new method of lossy still image compression, called Adap-tively Scanned Wavelet Difference Reduction (ASWDR). The ASWDR method produces an embedded bit stream with region of interest capability. It is a simple generalization of the compression method developed by Tian and Wells, which they have dubbed Wavelet Difference Reduction (WDR). While the WDR method employs a fixed ordering of the positions of wavelet coefficients, the ASWDR method employs a varying order which aims to adapt itself to specific image fea-tures. This image adaptive approach enables ASWDR to outperform WDR in a rate-distortion sense, and to essentially match the rate-distortion performance of the widely used codec, SPIHT, of Said and Pearlman. ASWDR compressed images exhibit better perceptual qualities, especially at low bit rates, than WDR and SPIHT compressed images. ASWDR retains all of the important features of WDR: low complexity, region of interest capability, embeddedness, and progres-sive SNR.
Article
Due to the increasing traffic caused by multimedia information and digitized form of representation of images; image compression has become a necessity. New algorithms for image compression based on wavelets have been developed. These methods have resulted in practical advances such as: superior low-bit rate performance, continuous-tone and bit-level compression, lossless and lossy compression, progressive transmission by pixel, accuracy and resolution, region of interest coding and others. We concentrate on the following methods of coding of wavelet coefficients, in this paper: EZW (embedded zero tree wavelet) algorithm, SPIHT (set partitioning in hierarchical trees) algorithm, SPECK (Set Partitioned Embedded Block Coder), WDR (wavelet difference reduction) algorithm, and ASWDR (adaptively scanned wavelet difference reduction) algorithm. These are relatively recent algorithms which achieve some of the lowest errors per compression rate and highest perceptual quality
Analysis based coding of image transform and sub band coefficients,‖ Applications of Digital Image Processing XVIII
  • V R Algazi
  • R R Estes
V. R. Algazi, R. R. Estes. -Analysis based coding of image transform and sub band coefficients,‖ Applications of Digital Image Processing XVIII, volume 2564 of SPIE Proceedings, pages 11-21, 1995.