Signal & Image Processing : An International Journal (SIPIJ) Vol.5, No.

6, December 2014

IMAGE COMPRESSION USING EMBEDDED

ZEROTREE WAVELET
A.M.Raid1, W.M.Khedr2, M. A. El-dosuky1 and Wesam Ahmed1
1

Mansoura University, Faculty of Computer Science and Information System

2
Zagazig University, Faculty of Science

ABSTRACT
Compressing an image is significantly different than compressing raw binary data. compressing images is
used by this different compression algorithm. Wavelet transforms used in Image compression methods to
provide high compression rates while maintaining good image quality. Discrete Wavelet Transform (DWT)
is one of the most common methods used in signal and image compression .It is very powerful compared to
other transform because its ability to represent any type of signals both in time and frequency domain
simultaneously. In this paper, we will moot the use of Wavelet Based Image compression algorithmEmbedded Zerotree Wavelet (EZW). We will obtain a bit stream with increasing accuracy from ezw
algorithm because of basing on progressive encoding to compress an image into . All the numerical results
were done by using matlab coding and the numerical analysis of this algorithm is carried out by sizing
Peak Signal to Noise Ratio (PSNR) and Compression Ratio (CR) for standard Lena Image .Experimental
results beam that the method is fast, robust and efficient enough to implement it in still and complex images
with significant image compression.

KEYWORDS
Discrete wavelet Transform, embedded zerotree wavelet, Image Compression

1. INTRODUCTION
Multimedia data needs considerable storage capacity and transmission bandwidth. There are
many forms of the data are the graphics, audio, video and image. [1]The main objective of image
compression is to reduce redundancy in image data in order to store or transmit only a minimal
number of samples from which a good approximation of the original image can be reconstructed
in accordance with human visual perception. In general, a text file can be compressed without any
errors up to original. This is called lossless compression. But after that extent errors are
unavoidable. A small loss in image compression is never noticeable. In lossy compression loss
and missed information is bearable. There are many types of lossless image compression include
variable-length encoding, Adaptive dictionary algorithms such as LZW, bit-plane coding, lossless
predictive coding, etc. Lossy compression has many classes include lossy predictive coding and
transform coding. Transform coding, which applies a Fourier-related transform like DCT and
Wavelet Transform like DWT are the most commonly used class. [2] Wavelet-based coding
supplies fundamental improvements in image quality at higher compression ratios. Over the past
few years, several competitive wavelet based Image Compression algorithms with an embedded
bit stream have been developed, such as Shapiros Embedded Zero tree Wavelet compression
(EZW) algorithm, Said and Pearlmans Set Partitioning In Hierarchical Trees (SPIHT) algorithm ,
Trumans Embedded Block Coding with Optimized Truncation (EBCOT) algorithm Wavelet
DOI : 10.5121/sipij.2014.5603

33

Signal & Image Processing : An International Journal (SIPIJ) Vol.5, No.6, December 2014

Difference Reduction (WDR) and Adaptively Scanned Wavelet Difference Reduction (ASWDR)
etc.[3][4] no many of research work has been published in literature on image compression
techniques using zerotree wavelet transform so this paper will contribute to the implementation of
the of lossy image compression technique (EZW algorithm).
The rest of the paper is organized as follows: section 2 explains image compression, section 3
explains (discrete wavelet transform (DWT)), section4 explains Embedded Zero tree Wavelet
compression (EZW) algorithm, section5 includes simulated results and section 5 gives the
conclusion and future work.

2. IMAGE COMPRESSION
Image Compression classifies the problem of reducing the amount of data required to represent
the digital image. The time is reduced by the reduction in file size required for image to be sent
over the internet or downloaded from web pages. Compression is achieved by the exit of one or
more of three basic data redundancies:
(1) Spatial Redundancy .
(2) Spectral redundancy .
(3) Psycho-visual redundancy .
The spatial and spectral redundancies are present because certain spatial and spectral patterns
between the pixels and the color components are common to each other, whereas the psychovisual redundancy originates from the fact that the human eye is insensitive to certain spatial
frequencies. [5]

2.1.Performance Criteria in Image Compression

The performance is rated by the following two Essential criteria: the obtained compression ratio
CR and the quality of the reconstructed image PSNR
(a) Compression ratio

CR = n1/n2

(1)

(b) Mean Square Error (MSE)

MSE =

(b)

1
HW

[ X (i, j) Y (i, j )]

(2)

i =1 j =1

PSNR

PSNR = 10 log10

255 2
(dB)
MSE

(3)

34

Signal & Image Processing : An International Journal (SIPIJ) Vol.5, No.6, December 2014

3. DISCRETE WAVELET TRANSFORMATION

The Wavelet Transform used to represent a time-frequency of the signal. It was developed to
cope the short coming of the Short Time Fourier Transform (STFT). The Wavelet Transform is
multi-resolution technique because the different frequencies are analyzed with different
resolutions.
Wavelet is one of lossy compression techniques have three procedures in general:
(1) Transform: we get a transformed data into wavelet domain. This step is invertible.
(2) Quantization: the wavelet coefficients are quantized to a limited alphabet. This step is not
Invertible.
(3) Entropy Coding: the resulting symbols after quantization are further compressed to
minimize the bit rate. This step is also invertible.
This simple diagram is a Wavelet Transform :

Figure 1. Wavelet transform diagram [6]

The Discrete Wavelet Transform (DWT) techniques have been resolved by Creusere (1997) to
decompose discrete time signals. The analysis scheme is named asub-band coding. Burt (1983)
determined a technique like sub-band coding and called it pyramidal coding which is also famed
as multi-resolution analysis. Later Vetterli et al (1990) made some improvements by removing
the existing redundancy in the pyramidal coding scheme. The basic idea of the DWT for a twodimensional image is termed as follows. An image is first decomposed into four parts based on
frequency sub bands, by critically sub sampling horizontal and vertical channels using sub band
filters and named as Low-Low (LL), Low-High (LH), High-Low (HL), and High- High (HH) sub
bands. To obtain the next coarser we will further decompose LL subband. This process is
repeated several times, figure 2 shows it.

Figure 2. Three scale wavelet decomposition [7]

35

Signal & Image Processing : An International Journal (SIPIJ) Vol.5, No.6, December 2014

The image of low frequency sub-band contains major information. High frequency sub-band
values approximate zero, the more high frequency the more obvious this situation. In the image,
the low frequency part acts the image information. So researchers take full advantage of the
characteristic after wavelet transform and employ proper method to process the image
coefficients for achieving further compression.

4. EZW ALGORITHM
The embedded zero tree wavelet algorithm (EZW) [7] is an effective image compression
algorithm. a fully embedded bit stream for image coding by this new technique. EZW algorithm
is an evident advantage. The user can select a bit rate and encode[8] the image to exactly the
wanted bit rate.
Features of EZW algorithm:
(1) Compact binary maps are provided by zero tree coding of significant wavelet coefficients.
(2) Successive approximation quantization of wavelet coefficients.
(3) Huffman coding.

4.1 Embedded Coding

The embedded zerotree wavelet algorithm (EZW) is a simple and efficient image compression
algorithm. [9] The bits in the bit stream are generated in order of importance in this coding
algorithm. [10] The schematic diagram of an makes the image transform, can be shown in Figure
3. The image coder has three main blocks: transform, quantizer, and entropy coder .

Figure 3. Embedded coding scheme [10]

In the case of the DWT (image transform), image is decomposed and producing the wavelet
coefficients at different scales. We use next block in Figure is a quantizer. It quantizes the
transformed matrix into a sequence of integers. After that using an entropy coding scheme.
In addition, The EZW encoder is based on two important annotations:
(1) The progressive encoding is a very normal option for compressing wavelet transformed
images, since the higher subbands only add detail.
(2) Large wavelet coefficients are more substantial than smaller wavelet coefficients. Shapiro
is a person who introduced Embedded zero-tree coding of wavelet coefficients (EZW).
In the Ezw algorithm if a wavelet coefficient at a coarse scale is insignificant with
ratingto a threshold T, then all wavelet coefficients of the same orientation in the same
spatial location at finer scale will be insignificant with respect To T. The trees induce a
parentchild relationship among the coefficients of subbands having the same spatial
36

Signal & Image Processing : An International Journal (SIPIJ) Vol.5, No.6, December 2014

orientation. These parentchild dependencies are at all events credited for the excellent
performance of zero-tree coders. Figure 4 shows The coarse scale coefficient is called the
parent and all coefficients corresponding to the same spatial location at the next finer
scale of similar orientation are called children .

Figure 4. Parent-child dependencies [10]

We summarize the EZW coding algorithm as follows: The EZW output stream will have to begin
with some information to synchronize the decoder. The minimum information required by the
decoder needs the number of wavelet transform levels used and the initial threshold, if we
suppose that always the same wavelet transform will be used. After imperfect reconstruction the
decoder can then replace the incomplete mean by the original mean. This can raise the PSNR
significantly. The EZW coding algorithm includes the first step to determine the initial threshold.
If we choose bitplane coding then our initial threshold t0 will be

(4)
MAX ((x,y)) is the maximum coefficient value in the image with this threshold. we enter the
main coding loop.
threshold = initial_threshold;
do {
dominant_pass(image);
subordinate_pass(image);
threshold = threshold/2;
} while (threshold > minimum_threshold);
(5)
We see that the image is coded by using two passes. the dominant pass is the first pass, a symbol
is outputted for every coefficient after scanning the image. If the coefficient is larger than the
threshold a P (positive) is coded, if the coefficient is smaller than minus the threshold an N
(negative) is coded. If the coefficient is the root of a zerotree then a T (zerotree) is coded and
finally, if the coefficient is smaller than the threshold but it is not the root of a zerotree, then a Z
(isolated zero) is coded. This happens when there is a coefficient larger than the threshold in the
tree. The effect of using the N and P codes is that when a coefficient is found to be larger than the
threshold (in absolute value or magnitude) its two most significant bits are outputted (if we forget
about sign extension).Note that in order to determine if a coefficient is the root of a zerotree or an
37

Signal & Image Processing : An International Journal (SIPIJ) Vol.5, No.6, December 2014

isolated zero, the subordinate pass is the second pass and is called the refinement pass. In this
list is ordered so that the largest coefficients are again transmitted first. The main loop stops when
the threshold reaches a minimum value. For integer coefficients is minimum value equals zero
and the divide by two and after that can be replaced by a shift right operation. If we append
another ending condition based on the number of outputted bits by the huffman coder then we can
apply any target bit rate exactly without doing too much work.
We abstracted the above with the following code fragments, starting with the dominant pass. In
the dominant_pass as shown in figure 5:

Figure 5. Flowchart for dominant pass [10]

In the dominant_pass : all the coefficients are larger than the current threshold are extracted and
placed without their sign on the subordinate list after that we will make their positions in the
image are filled with zeroes. This will make them from being coded still.
In the subordinate_pass : All the coefficients in the subordinate list are refined. this gives boost to
some juggling with uncertainty intervals and it products next most significant bit of all the
coefficients in the subordinate list
.

5. SIMULATED RESULTS
We stratified the embedded zero-tree wavelet algorithm (EZW) on the lena image. The input
image file is used in BMP format. The image size was considered to be 512*512 pixels. , the
image was compressed and reconstructed by using a threshold value of 8 for EZW algorithm and
using Huffman coding. The compression ratio and the calculated PSNR value determined for the
image Lena in table 1.

Figure 6. Original image

Figure 7. EZW reconstructed image

38

Signal & Image Processing : An International Journal (SIPIJ) Vol.5, No.6, December 2014
Table 1. Performance evaluation of dwt-ezw algorithm

Technique
Ezw

CR
1.6647

PSNR(dB)
30. 46

6. CONCLUSION & FUTURE WORK

This paper is about Embedded Zerotree Wavelet (EZW) algorithm which is very much indeed
effective and important compression algorithm. This algorithm is invent by Shapiro in (1993) and
it is an image compression algorithm that the bits in the bit stream are produced in order of
importance. The algorithm is dependent on the selection of significant coefficients with respect to
given thresholds. In this paper, we tried to use the embedded zero tree wavelet compression on
color Lena image and the PSNR and compression ratio values fixed for the Image.
In the future work we will make a comparison between EZW (Embedded Zero tree wavelet) and
SPIHT (Set Partitioning in Hierarchical Trees) coding techniques in color image compression.

REFERENCES
[1]

R.C. Gonzalez, R. E. Woods, and S..L. Eddins.Digital Image Processing Using MATLAB.. ISBN10:0130085197. ISBN-13: 978-0130085191. Prentice Hall, 1st edition ,September 5, 2003.
[2] P. Singh, Nivedita and S.Sharma, A Comparative Study: Block Truncation Coding, Wavelet,
Embedded Zerotree and Fractal Image
Compression on ColorImage, International Journal of
Electronics and Communication Engineering & Technology(IJECET), Volume 3, Issue 2, ISSN Print:
0976- 6464, ISSN Online:0976 6472, pp. 10 - 21,2012.
[3] R. A. DeVore, B. Jawerth, and B. J. Lucier, Image Compression Through Wavelet Transform
Coding, IEEE Trans. On Information Theory, Vol. 38, No. 2,pp. 719 - 746, 1992.
[4] H. J. Barnard, Image and Video Coding Using Wavelet Transforms, Ph.D. Dissertation, Technische
Universiteit Delft, May 1994.
[5] N. Ahmed, T. Natarajan, and K. R. Rao, Discrete Cosine Transform, IEEE. Trans. Computer,
Vol
C-23, pp. 90-93 Jan 1974.
[6] S. Pemmaraju and S. Mitra, Efficient image coding using multiresolution wavelet transform and
vector quantization IEEE Southwest Symposium on Image Analysis and Interpretation, IEEE press,
Santa Fe, New Mexico, pp:243-248 ,April, 1996.
[7] J.M.Shapiro,Embedded Image Coding Using Zerotrees of Wavelet Coefficients, IEEE Trans. on
Signal Processing, Vol. 41, pp 3445-3462, 1993.
[8] R. Ponalagusamy and C .Saravanan," Analysis of MedicalImage Compression using Statistical
Coding Methods, Advances inComputer Science and Engineering: Reports and Monographs,Imperial
College Press, UK, Vol.2., pp 372-376, 2007.
[9] A. S. Lewis and G. Knowles, Image Compression Using the 2-D Wavelet Transform, IEEE Trans.
On Image Processing, Vol. 1, No. 2, pp. 244-250, April, 1992.
[10] V.S.Shingate, T.R. Sotakke and S.N.Talbar, Still Image Compression using Embedded Zerotree
Wavelet Encoding, Inrenational Journal of Computer Science & Communication, Vol.1,No.1,,pp.2124 January-June 2010.

39