Comparative Analysis of Various Wavelet Filters For Compression of An Image

International Journal of Advanced Engineering Research and Technology (IJAERT), ISSN: 23488190
ICRTIET-2014 Conference Proceeding, 30th -31st August 2014
Comparative Analysis of Various Wavelet Filters for Compression of

An Image
Kamya Gupta1, Mohit Lalit2
1
2
Department of Computer Science, Kurukshetra University, Kurukshetra

Department Of Computer Science, Kurukshetra University, Kurukshetra
Abstract
The rapid development of information and
communication technologies is enabling large amount of
information to be processed, stored, and transmitted over
high speed networks. The need for data compression and
transmission is increasingly becoming a significant topic
in all areas of computing and communications.
Computing techniques that would considerably reduce
the image size that occupies less space and bandwidth
for transmission over networks form an active research.
Image compression deals with reducing the amount of
data required to represent a digital image. Uncompressed
multimedia (graphics, audio and video) data requires
considerable storage capacity and transmission
bandwidth. Despite rapid progress in mass-storage
density, processor speeds, and digital communication
system performance, demand for data storage capacity
and data-transmission bandwidth continues to outstrip
the capabilities of available technologies. The problem
inherent to any digital image is the large amount of
bandwidth required for transmission or storage. Artificial
neural networks are becoming attractive in image
processing where high computational performance and
parallel architectures are required. The good picture
quality can be retrieved if Wavelet-based image
compression technique is used for compression and also
provides better compression ratio.
Here, we are going to study in detail the image
compression techniques, implement wavelet based
compression algorithm, applying various wavelet filters
to get comparative analysis in MSE, PSNR, Retained
Energy in the compressed images.
Keywords Compression, PSNR, Retained Energy,
Wavelet Filter.
1. Introduction
Uncompressed multimedia (graphics, audio and video)
data requires considerable storage capacity and
transmission bandwidth. The recent growth of data
intensive multimedia-based web applications has not
only sustained the need for more efficient ways to
storage and communication technology. Image
compression addresses the problem of reducing the
amount of data required to represent a digital image.
An image can be mathematically defined as a two-
dimensional function f (x,y) where x and y are spatial

coordinates. Given an image of dimensions MN,
each pair of coordinates (x,y) defines an element in
the image called a pixel. The amplitude of f for any
pixel is the intensity of the image at that point. If the
coordinate and amplitude values are all finite and
discrete, then the image is called a digital image.
Digital image processing can be classified broadly into
four areas:
I. Image Enhancement
I. Image Restoration
II. Image Coding
III. Image Understanding
When we say processed images, we mean images
captured by camera or other scanned images. Image
processing has been in place for a long period of time.
These images can be taken in different domain i.e. snap
shots, scanned documents etc, which should need some
modification for a clearer look at it and this is why image
processing is in place.
The principles of image compression are based on
information theory. The amount of information that a
source produces is called Entropy. The amount of
information one receives from a source is equivalent to
the amount of the uncertainty that has been removed. A
source produces a sequence of variables from a given
symbol set. For each symbol, there is a product of the
symbol probability and its logarithm. The entropy is a
negative summation of the products of all the symbols in
a given symbol set. Encoding and Decoding during
Image Compression.
Wavelets are functions used in representing data or other
functions. They analyze the signal at different
frequencies with different resolutions. They represents
optimized standard for images with amount of data with
sharp discontinuities. Wavelet compression transforms
the entire image differently from JPEG and is more
natural as if follows the shape of the objects in the
picture. It is necessary to use special software for
viewing, being this a non-standardized compression
method. Wavelet analysis can be used to divide the
information of an image into approximation and detail
sub-signals. The approximation sub-signal shows the
general trend of pixel values, and three detail sub-signals
Divya Jyoti College of Engineering & Technology, Modinagar, Ghaziabad (U.P.), India
16
show the vertical, horizontal and diagonal details or

changes in the image. If these details are very small then
they can be set to zero without significantly changing the
image. The value below which details are considered
small enough to be set to zero is known as the threshold.
The greater the number of zeros the greater the
compression that can be achieved. The amount of
information retained by an image after compression and
decompression is known as the Energy Retained and this
is proportional to the sum of the squares of the pixel
values.If the energy retained is 100% then the
compression is known as lossless, as the image can be
reconstructed exactly. This occurs when the threshold
value is set to zero, meaning that the detail has not been
changed. If any values are changed then energy will be
lost and this is known as lossy compression. Ideally,
during compression the number of zeros and the energy
retention will be as high as possible. However, as more
zeros are obtained more energy is lost, so a balance
between the two needs to be found.
1.1 Artificial Neural Network

An Artificial Neural Network (ANN) is an informationprocessing model that is inspired by the way biological
nervous systems, such as the brain, process information.
The key element of this model is the new structure of the
information processing system. It is composed of a large
number of highly interconnected processing elements
(neurons) working in collectively to solve specific
problems. Unsupervised system are basically based on
competitive learning, in which the output neurons
compete amongst themselves to be activated, with the
result that only one is activated at any one time. This
activated neuron is called a winner-takes-all neuron or
simply the winning neuron. Such competition can be
induced/ implemented by having lateral inhibition
connections (negative feedback paths) between the
neurons. The result is that the neurons are forced to
organize themselves. For obvious reasons, such a
network is called a Self Organizing Map (SOM).
1.2 MatLab
Matlab is designed to solve problems numerically, that
is, in finite-precision arithmetic. Therefore it produces
approximate rather than exact solutions, and should not
be confused with a symbolic computation system (SCS)
such as Mathematica or Maple. It should be understood
that this does not make Matlab better or worse than an
SCS; it is a tool designed for different tasks and is
therefore not directly comparable. In 2004, MATLAB
had around one million users across industry and
academia. MATLAB users come from various
backgrounds of engineering, science, and economics.
MATLAB is widely used in academic and research
institutions as well as industrial enterprises. MATLAB is
also a programming language. By creating a file with the

extension .m we can easily write and run programs. If
one has to create a program file myfile.m in the
MATLAB language, then one can make the command
myfile from MATLAB and it will run like any other
MATLAB function. There is no need to compile the
program since MATLAB is an interpretative (not
compiled) language. Such a file is called an m-file.
Wavelet Toolbox software is a collection of functions
built on the MATLAB technical computing
environment. It provides tools for the analysis and
synthesis of signals and images, and tools for statistical
applications, using wavelets and wavelet packets within
the framework of MATLAB.
1.2.1 Wavelet Applications
Wavelets have scale aspects and time aspects;
consequently every application has scale and time
aspects. For scale aspects, one idea is presented
around the notion of local regularity. For time
aspects, lists of domains are presented. When the
decomposition is taken as a whole, the de-noising
and compression processes are center points. Many
applications use the wavelet decomposition taken as
a whole. The common goals concern the signal or
image clearance and simplification, which are parts
of de-noising or compression.
1.2.2 Wavelet Families
Several families of wavelets that have proven to be
especially useful are included in this toolbox. What
follows is an introduction to some wavelet families.
o Haar
o Daubechies
o Biorthogonal
o Coiflets
o Symlets
o Morlet
o Mexican Hat
o Meyer
o Other Real Wavelets
o Complex Wavelets
1.2.3 Wavelet Analysis
Wavelet analysis represents the next logical step: a
windowing technique with variable-sized regions.
Wavelet analysis allows the use of long time
intervals where we want more precise low-frequency
information, and shorter regions where we want
high-frequency information.
2. Literature Review
A brief review of the work done in the field is
described here:
17
Abdul Khader Jilani Saudagar [1] implements an

enhanced image coding system for biomedical image
compression compared to the existing JPEG2000
system. It is observed that proposed algorithm is able
to achieve good quality performance with a
relatively simple algorithm. Proposed algorithm does
not require complicated bit allocation procedures
like subband coding does, and it does not require
prior knowledge of the image source like JPEG,
JPEG2000 does (to optimize quantization tables).
The results obtained from hybrid neural networks
found much better results when compared to
conventional JPEG2000 approach.
Adnan Khashman, Kamil Dimililer [2] proposed a
neural network could be trained to recognize an
optimum ratio for Haar wavelet compression of an
image upon presenting the image to the network.
They developed two neural networks receiving
different input image sizes and a comparison
between their performances in finding optimum
Haar-based compression.
Anuj Bhardwaj and Rashid Ali [3] describe the
algorithm for image compression using MFHWT
and shows better results than those obtained by using
any other method on an average. It includes a
number of examples of different images to validate
the utility and significance of algorithms
performance.
A. Dwivedi et. al. [4] proposed a novel hybrid image
compression technique. This technique inherits the
properties of localizing the global spatial and
frequency
correlation
from
wavelets
and
classification and functional approximation tasks
from modified forward-only counter propagation
neural network (MFOCPN) for image compression.
G. Boopathi and S. Arockiasamy [5] proposed a
technique for image compression using modified
Self-Organizing Map (SOM) based vector
quantization. Self-Organizing Feature Map (SOFM)
algorithm is a type of neural network model which
consists of one input and one output layer. Each
input node is connected with output node by
adaptive weights. The experimental result shows that
the proposed technique obtained better PSNR value
end also reduces Mean Square Error.
H. S. Kong, L. Guan [6] proposed a new method for
multiscale image segmentation. The method is based
on pixel classification by means of a self-organizing
neural network. The core concept of this processing
method is to explicitly treat segmentation as a
classification problem. An unsupervised learning

algorithm is utilized in the processing. Compared
with other segmentation methods, the proposed one
has a number of desirable features. It is selfadaptive, efficient, and easy to control. The
effectiveness of the proposed method is verified
through several experiments.
K. Siva Nagi Reddy et. al. [7] presented a neural
network based image compression method. Neural
networks offer the potential for providing a novel
solution to the problem of data compression by its
ability to generate an internal data representation.
This network, which is an application of back
propagation network, accepts a large amount of
image data, compresses it for storage or
transmission, and subsequently restores it when
desired. A new approach for reducing training time
by reconstructing representative vectors has also
been proposed. Performance of the network has been
evaluated using some standard real world images.
Kamrul Hasan et. al. [8] presented a low complex
2D image compression method using Haar wavelets
as the basis functions along with the quality
measurement of the compressed images. The 2D
discrete wavelet transform (DWT) has been applied
and the detail matrices from the information matrix
of the image have been estimated. The reconstructed
image is synthesized using the estimated detail
matrices and information matrix provided by the
Wavelet transform. The quality of the compressed
images has been evaluated using some factors like
Compression Ratio (CR), Peak Signal to Noise Ratio
(PSNR), Mean Opinion Score (MOS), Picture
Quality Scale (PQS) etc. Haar Wavelet is chosen for
implementation.
M. Mozammel et. al. [9] proposed a new image
compression scheme based on discrete wavelet
transform which provides
sufficient
high
compression ratios with no appreciable degradation
of image quality. The effectiveness and robustness of
this approach has been justified using a set of real
images.
To demonstrate the performance of the proposed
method, a comparison between the proposed
technique and other common compression
techniques has been revealed. Wavelets are better
suited to time-limited data and wavelet based
compression technique maintains better image
quality by reducing errors.
Madhuri Khambete and Madhuri Joshi [10] proposed
a novel method to quantify blur and ringing artifact
18
in an image. Quality evaluation of an image is an

important task in image processing applications. In
case of image compression, quality of decompressed
image is also the criterion for evaluation of given
coding scheme. In the process of compression
decompression various artifacts such as blocking
artifacts, blur artifact, ringing or edge artifact are
observed. Quantification of these artifacts is a
difficult task.
3. Proposed Work
The area of image compression is well established
and a lot of work has been done. However, there is a
need of high accuracy and less training time is
always required. Most of the research work done
lacks in the above constraints. Hereby we are going
to study the image compression in detail using
various methods by giving a brief literature review.
A detailed study about the Wavelet based Image
compression is presented. Hereby, we are going to
present and compare various wavelet filters for the
image compression.
Two measures are commonly used to evaluate the
perceptual quality:
The Mean Square Error (MSE). It represents the
mean squared error between the compressed and the
original image and is given by:
Step 4: A wavelet filters is applied on the image.

There are various type of wavelet filters are
available. Syntax for applying wavelet filter is:
[Lo_D,Hi_D,Lo_R,Hi_R] = wfilters('wname')
Step 5: Multilevel 2-D wavelet decomposition is
achieved using wavedec2() whose syntax is given
below:
[C,S] = wavedec2(X,N,Lo_D,Hi_D)
wavedec2 is a two-dimensional wavelet analysis
function.
Step 6: Thresholds for wavelet is determined using
wdcbm2() whose syntax is given as:
[THR,NKEEP] = wdcbm2(C,S,ALPHA,M)
Step 7: Compression is done using wdencmp() and
the syntax is:
[XC,CXC,LXC,PERF0,PERFL2]
=
wdencmp('lvd',C,L,'wname',N,THR,SORH)
Step 8: Reconstruction of Multilevel 2-D wavelet is
done using waverec2()
X = waverec2(C,S,'wname')
Where waverec2 is a two-dimensional wavelet
analysis function.
Step 9: Compressed Image is displayed. Various
outcomes like Retained Energy and PSNR are
computed for the particular wavelet filter.
The lower the value of MSE, the lower the error.

The Peak Signal to Noise Ratio (PSNR). It
represents a measure of the peak error and is
expressed in decibels. It is defined by:
The higher the PSNR, the better the quality of the

compressed or reconstructed image.
Implementation
Here, we are illustrating the present work done to
achieve the objective of compression using wavelet
method in the form of steps.
Step 1: Bitmap Image is read/loaded using imread
function.
Step 2: A check for the size of image is made that
will check the size in the form of dimensions and
will resize the image as per the best conditions.
Step 3: A decomposition level is taken from the user
for decomposing the image.
4. Result Analysis
Here, Input Image taken is of size 2.25 MB at which
various wavelet filters with different decomposition
levels are applied. Initially, we are going to represent
Compression Ratio, Retained Energy, PSNR and the
image file size after the compression at a particular
decomposition level for wavelet filter db1, db2. In
the same way values for other filters are obtained.
Afterwards, we are showing collective results of all
the filters chosen for our research work with
compression parameter values obtained after
applying wavelet filter with particular decomposition
level.
Wavelet Filter = db1
For the decomposition level=1, the compression
ratio: 13.1201, Retained Energy: 99.9963, PSNR
value: 103.9338, Reconstructed Image Size=101
KB.
value: 85.0442, Reconstructed Image Size=91 KB.
19
A comparison graph for the Compression Ratio,

Retained Energy, PSNR and Image Size for
decomposition level=1, 2, 3 is shown in figure 1.
Fig. 1: db1 filter attributes for decomposition

level=1,2,3
Wavelet Filter = db2
ratio: 6.911, Retained Energy: 99.9972, PSNR value:
100.2954, Reconstructed Image Size=101 KB.
A comparison graph for the Compression Ratio,
Retained Energy, PSNR and Image Size for
decomposition level=1, 2, 3 is shown in figure 2.
Compression factor values obtained after applying

the implementation for decomposition level=2 are
shown in figure 4.
Fig. 4: Graph showing comparison at decomposition

level 2 of various filters
shown in figure 5.
,

Fig. 2: db2 filter attributes for decomposition
level=1,2,3
shown in figure 3.

5. Conclusion
A detailed literature survey of the image compression
using various compression methods e.g. Wavelet, DCT,
Neural Network etc. Also work done in the field using
Wavelet Analysis has been reviewed in detail. Hereby,
we have used Wavelet based compression methods for
image compression using various wavelet filters at
different decomposition levels. We have studied various
image processing methods, wavelet functions and filters
to apply for the image compression problem. The results
taken from the MatLab model suggests that this
algorithm has the ability to compress images of any size
to much lesser than the original size and shows the
image quality can be maintained with keeping
decomposition level and wavelet filter of choice
depending on the criteria of PSNR, Compression ratio,
Retained Energy etc.
20
6. Limitations & Future Scope

There are various Image compression methods available.
Major limitation is to accommodate images of very large
size. Also many improvements in the result can be
finding out by applying other methodologies. The work
can be extended by applying Discrete Cosine
Transformation (DCT), Soft Computing techniques like
Neural Network (Supervised/Unsupervised), Fuzzy,
Genetic methods etc. A detailed analysis can be made
that for what scenario which methodology will provide
optimum results.
References
[1] Abdul Khader Jilani Saudagar, Omar A. Shathry,
Neural Network Based Image Compression
Approach to Improve the Quality of Biomedical
Image for Telemedicine, British Journal of Applied
Science & Technology, Vol. 4, No. 3, pp. 510-524,
2014.
[2] Adnan Khashman, Kamil Dimililer, Image
Compression using Neural Networks and Haar
Wavelet, WSEAS Transactions On Signal
Processing, Vol. 4, Issue 5, pp. 330-339, May 2008.
[5] G. Boopathi, S. Arockiasamy, An Image

Compression Approach using Wavelet Transform
and Modified Self Organizing Map, IJCSI
International Journal of Computer Science Issues,
Vol. 8, Issue 5, No. 2, pp. 323-330, Sept. 2011.
[6] H. S. Kong, L. Guan, S. Y. Kung, A self organizing
tree map approach for image segmentation, Proc. of
the 6th IEEE International Conference on Signal
Processing, pp. 588-591, 2002.
[7] K.Siva Nagi Reddy, B.R.Vikram, L. Koteswara
Rao, B.Sudheer Reddy, Image Compression and
Reconstruction Using a New Approach by Artificial
Neural Network, International Journal of Image
Processing (IJIP), Vol. 6, Issue 2, pp. 68-85, 2012.
.
[8] Kamrul Hasan Talukder, Koichi Harada, Haar
Wavelet Based Approach for Image Compression
and Quality Assessment of Compressed Image,
IAENG
International
Journal
of
Applied
Mathematics, Vol. 36, Issue 1, Feb. 2007.
[3] Anuj Bhardwaj, Rashid Ali, Image Compression

Using Modified Fast Haar Wavelet Transform,
World Applied Sciences Journal, Vol. 7, No. 5, pp.
647-653, 2009.
[9] M. Mozammel Hoque Chowdhury, Amina Khatun,

Image Compression Using Discrete Wavelet
Transform, IJCSI International Journal of Computer
Science Issues, Vol. 9, No. 1, Issue 4, pp. 327-330,
July 2012
[4] A. Dwivedi, N. S. C. Bose, A. Kumar, P. Kandula,

D. Mishra, P. K. Kalra, A Novel Hybrid Image
Compression
Technique:
Wavelet-MFOCPN,
Proceedings of the 9th Asian Symposium on
Information Display, pp. 492-495, Oct. 2006.
[10]
Madhuri Khambete, Madhuri Joshi, Blur and
Ringing
Artifact
Measurement
in
Image
Compression using Wavelet Transform, World
Academy of Science, Engineering and Technology
2, pp. 183-186, 2007.
21

Comparative Analysis of Various Wavelet Filters For Compression of An Image

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Comparative Analysis of Various Wavelet Filters For Compression of An Image

Uploaded by

Copyright:

Available Formats

International Journal of Advanced Engineering Research and Technology (IJAERT), ISSN: 23488190

ICRTIET-2014 Conference Proceeding, 30th -31st August 2014

Comparative Analysis of Various Wavelet Filters for Compression of

Department of Computer Science, Kurukshetra University, Kurukshetra

dimensional function f (x,y) where x and y are spatial

show the vertical, horizontal and diagonal details or

1.1 Artificial Neural Network

also a programming language. By creating a file with the

Abdul Khader Jilani Saudagar [1] implements an

classification problem. An unsupervised learning

in an image. Quality evaluation of an image is an

Step 4: A wavelet filters is applied on the image.

The lower the value of MSE, the lower the error.

The higher the PSNR, the better the quality of the

A comparison graph for the Compression Ratio,

Fig. 1: db1 filter attributes for decomposition

Compression factor values obtained after applying

Fig. 4: Graph showing comparison at decomposition

Fig. 5: Graph showing comparison at decomposition

Fig. 3: Graph showing comparison at decomposition

6. Limitations & Future Scope

[5] G. Boopathi, S. Arockiasamy, An Image

[3] Anuj Bhardwaj, Rashid Ali, Image Compression

[9] M. Mozammel Hoque Chowdhury, Amina Khatun,

[4] A. Dwivedi, N. S. C. Bose, A. Kumar, P. Kandula,

You might also like