Sei sulla pagina 1di 5

International Journal of Emerging Technology and Advanced Engineering

Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 4, Issue 3, March 201.5

Discrete Wavelet Transform for Image Processing


Dipalee Gupta1, Siddhartha Choubey2
Abstract— Image compression is a method through which In order to compress the image, redundancies must
we can reduce the storage space of images which will helpful be exploited, for example, areas where there is little or no
to increase storage and transmission process’s performance. change between pixel values. Therefore images having
In this paper, we present the comparison of the performance large areas of uniform colour will have large redundancies,
of discrete wavelets like Haar Wavelet and Daubechies
and conversely images that have frequent and large
Wavelet for implementation in a still image compression
system. The performances of these transforms are compared changes in colour will be less redundant and harder to
in terms of Mean squared error (MSE) and Energy Retained compress[3]. The objective of image compression is to
(ER) etc. The main objective is to investigate the still image reduce redundancy (i.e. Coding redundancy, Interpixel
compression of a gray scale image using wavelet theory. This redundancy & psycho visual redundancy) of the image
is implemented in software using MATLAB Wavelet Toolbox data in order to be able to store or transmit data in an
and 2D-DWT technique. The experiments and results is efficient form. Data compression is achieved when one or
carried out on .jpg format images. These results provide a more of these redundancies are reduced or eliminated.
good reference for application developers to choose a good There are several methods of image compression
wavelet compression system for their application.
available today that is lossless and lossy image
Keywords—Image Compression, Discrete Wavelet compression. In lossless compression , every single bit of
Transform, Haar Wavelet, Daubechies Wavelet , Wavelet data that was originally in the file remains after the file is
Toolbox (WT) uncompressed. All of the information is completely
restored. Medical imaging, technical drawings, and
I. INTRODUCTION astronomical observations typically use lossless
compression techniques. The Graphics Interchange File
Compression is one of the major image processing
(GIF) is an image format used on the Web that provides
techniques. It is one of the most useful and commercially
lossless compression. In lossy compression reduces a file
successful technologies in the field of digital image
by permanently eliminating certain information, especially
processing. Image compression is the representation of an
redundant information. When the file is uncompressed,
image in digital form with as few bits as possible while
only a part of the original information is still there. Pictures
maintaining an acceptable level of image quality[1].
and videos from digital cameras are examples of digital
Increasingly images are acquired and stored digitally or
files that are commonly compressed using lossy methods.
various film digitizers are used to convert traditional raw
A simple method of lossy image compression is to reduce
images into digital format. Data$ compression is the
the color space to a smaller set of colors. The JPEG image
technique to reduce the redundancies in data representation
file, commonly used for photographs and other complex
in order to decrease data storage requirements and hence
still images on the Web, is an image that has lossy
communication costs. Reducing the storage requirement is
compression[4-5].
equivalent to increasing the capacity of the storage medium
Wavelets have been first found in the literature in works
increase the speed of transmission and hence
of Grossmann and Morlet. Wavelet analysis is a
communication bandwidth[2]. The efficient ways of storing
mathematical technique used to represent data or
large amount of data and due to the bandwidth and storage
functions[6]. The wavelet transform is an emerging signal
limitations, images must be compressed before
processing technique that can be used to represent real-life
transmission and storage.At some later time, the
non- stationary signals with high efficiency. A wavelet is a
compressed image is decompressed to reconstruct the
waveform of effectively limited duration that has an
original image or approximation of it.
average value of zero. The term ‗wavelet‘ comes from the
Redundancy is that portion of data that can be removed
fact that they integrate to zero; they wave up and down
when it is not needed or can be reinserted to interpret the
across the axis. Many wavelets also display a property ideal
data when needed. Most often, the redundancy is reinserted
for compact signal representation: orthogonality[7].
in order to generate the original data in its original form.
An image can be thought of as a matrix of pixel values.

598
International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 4, Issue 3, March 201.5
Discrete wavelet transforms is the most popular The signal to be analyzed is passed through filters with
transformation technique adopted for image compression. different cut-off frequencies at different scales. It is easy to
So the proposed methodology of this paper is to achieve implement and reduces the computation time and resources
high compression ratio in images through implementing required[9]. A 2-D DWT can be seen as a 1-D wavelet
Haar Wavelet Transform and daubachies wavelet scheme which transform along the rows and then a 1-D
transform using software tools MATLAB. Through wavelet transform along the columns,. The 2-D DWT
experimental results we compare performance of both the operates in a straight forward manner by inserting array
transforms. We works in the MATLAB computing transposition between the two 1-D DWT. The rows of the
environment, the Wavelet Toolbox offers some significant array are processed first with only one level of
advantages. The amount of information retained by an decomposition. This essentially divides the array into two
image after compression and decompression is known as vertical halves, with the first half storing the average
the energy retained, and this is proportional to the sum coefficients, while the second vertical half stores the detail
of the squares of the pixel values. If the energy retained is coefficients. This process is repeated again with the
100% then the compression is known as lossless. This columns, resulting in four sub-bands (see Fig. 1) within the
occurs when the threshold value is set to zero, meaning that array defined by filter output. Fig. 1 shows a three- level 2-
the detail has not been changed. If any values are changed D DWT decomposition of the image.
then energy will be lost and this is known as lossy
compression. Ideally, during compression the no 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. This
article is structured as follows: Section 2 describes the
Discrete Wavelet Transform, Section 3 illustrates the
Methodology , Section 4 presents the result and discussion.
Finally, Section 6 concludes the paper.

II. DISCRETE WAVELET TRANSFORM


The wavelet transform has gained widespread
acceptance in signal processing and image compression.
Recently the JPEG committee has released its new image
coding standard, JPEG-2000, which has been based upon
DWT. Wavelet transform decomposes a signal into a set of
Fig 1. Three level decomposition for 2D – DWT
basis functions. These basis functions are called wavelets.
Wavelets are obtained from a single prototype wavelet Image consists of pixels that are arranged in two
called mother wavelet by dilations and shifting[8]. The dimensional matrix, each pixel represents the digital
DWT has been introduced as a highly efficient and flexible equivalent of image intensity. In spatial domain adjacent
method for sub band decomposition of signals. The 2D- pixel values are highly correlated and hence redundant. In
DWT is nowadays established as a key operation in image order to compress images, these redundancies existing
processing .It is multi-resolution analysis and it among pixels needs to be eliminated. DWT processor
decomposes images into wavelet coefficients and scaling transforms the spatial domain pixels into frequency domain
function. In Discrete Wavelet Transform, signal energy information that are represented in multiple sub-bands,
concentrates to specific wavelet coefficients. This representing different time scale and frequency points. One
characteristic is useful for compressing images[9]. of the prominent features of JPEG2000 standard, providing
Wavelets convert the image into a series of wavelets it the resolution scalability , is the use of the 2D-DWT to
that can be stored more efficiently than pixel blocks. convert the image samples into a more compressible form.
Wavelets have rough edges, they are able to render pictures The JPEG 2000 standard proposes a wavelet transform
better by eliminating the ―blockiness‖. In DWT, a time- stage since it offers better rate/distortion (R/D)
scale representation of the digital signal is obtained using performance than the traditional DCT.
digital filtering techniques.

599
International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 4, Issue 3, March 201.5
A. Haar Transform
Haar functions have been used from 1910 when they
were introduced by the Hungarian mathematician Alfred
Haar[10]. Haar wavelet is discontinuous, and resembles a
step function. It represents the same wavelet as Daubechies
db1. Haar used these functions to give an example of an
orthonormal system for the space of square-integrable
function on the unit interval [0, 1].
Fig 3. Daubechies Wavelet

The Daubechies wavelet transforms are defined in the


same way as Haar wavelet transform by computing the
running averages and differences via scalar products with
scaling signals and wavelets the only difference between
them consists in how these scaling signals and wavelets are
defined[13]. This wavelet type has balanced frequency
responses but non-linear phase responses. Daubechies
wavelets use overlapping windows, so the high frequency
coefficient spectrum reflects all high frequency changes.
Therefore Daubechies wavelets are useful in compression
Fig 2. Haar Wavelet and noise removal of audio signal processing.
For an input represented by a list of numbers, the Haar
wavelet transform may be considered to simply pair up III. METHODOLOGY
input values, storing the difference and passing the sum. Following steps are performed for compression :
This process is repeated recursively, pairing up the sums to
a) Load the image which is compressed.
provide the next scale, finally resulting in differences and
b) Applying the transform-The compression
one final sum. The Haar Wavelet Transformation is a
algorithm starts by transforming the image from
simple form of compression which involves averaging and
data space to wavelet space. This is done on
differencing terms, storing detail coefficients, eliminating
several levels.
data, and reconstructing the matrix such that the resulting
c) Chossing the threshold- neglect all the wavelet
matrix is similar to the initial matrix.[11-12].
coefficients that fall below a certain threshold. We
A Haar wavelet is the simplest type of wavelet. In
select our threshold in such a way as to preserve a
discrete form, Haar wavelets are related to a mathematical
certain percent of the total coefficients - this is
operation called the Haar transform. The Haar transform
known as ‖quantile‖ thresholding.
serves as a prototype for all other wavelet transforms. Like
d) Perform compression at different transform
all wavelet transforms, the Haar transform decomposes a
discrete signal into two sub-signals of half its length. One
sub-signal is a running average or trend; the other sub- IV. RESULT
signal is a running difference or fluctuation. In this paper, we compared Haar and Daubechies
wavelet of Discrete wavelet transform (DWT). The quality
B. Daubechies Transform of a compression method could be measured by the
Ingrid Daubechies, one of the brightest stars in the world traditional distortion measures such as Mean square error
of wavelet research, invented what are called compactly (MSE) and energy retained(ER). We compared the
supported orthonormal wavelets - thus making discrete performance of these transforms on image
wavelet analysis practicable. The names of the Daubechies ―Koala.jpg‖(512x512).
family wavelets are written dbN, where N is the order, and
db the ―surname‖ of the wavelet. The db1 wavelet, as
mentioned above, is the same as Haar wavelet. Here is the
wavelet functions psi of the next nine members of the
family.

600
International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 4, Issue 3, March 201.5

Fig 4. Original Image (b)


Fig 6. Compressed Imge (a) Haar (b) Daubechies

(a)

.
Fig 5. Decomposition al level 3

(b)
Where

(a)
Fig 7. Graph shows the retained energy and zeros in %
(a) Haar Wavelet (b) Daubechies Wavelet

601
International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 4, Issue 3, March 201.5
In lossy compression the information is loss, so we not
restore all the information. Ideally, during compression the
no of zeros and the energy retention will be as high as
possible. In Daubechies wavelet retained energy is 98.03%
and in haar transform it is 97.70%.
So from the results, it has been concluded that
Daubechies wavelet transform shows the best results in
terms of Energy Retain as compare to Haar wavelet
transform.
REFERENCES
[1] Priyanka Singh, Priti Singh, Rakesh Kumar Sharma, 2011, ―JPEG
(a) Image Compression based on Biorthogonal, Coiflets and Daubechies
Wavelet Families‖ International Journal of Computer Applications ,
Volume 13– No.1.
[2] Maneesha Gupta, Dr.Amit Kumar garg, Mr.Abhishek Kaushik,
2011, ―Review: Image Compression Algorithm‖ IJCSET , Vol 1.
[3] B.B.S.Kumar, Dr.P.S.Satyanarayana , 2013, ―Image Analysis Using
Biorthogonal Wavelet‖ Published in International Journalof
Innovative Research And Development, Vol 2.
[4] Ms. Sonam Malik and Mr. Vikram Verma , 2012, ―Comparative
analysis of DCT, Haar and Daubechies Wavelet for Image
Compression‖ Published in International Journal of Applied
Engineering Research, Vol.7-No.11.
[5] Sanjeev Singla, Abhilasha Jain, 2013, ―Improved 2-D DCT Image
Compression Using optimal compressed value‖ Sanjeev et al. /
(b) IJAIR, Vol. 2.
[6] Daljeet Kaur Khanduja & M.Y.Gokhal ―Time Domain Signal
Analysis Using Modified Haar and Modified Daubechies Wavelet
Transform‖ Signal Processing-An International Journal (SPIJ),
Volume (4).
[7] M. Mozammel Hoque Chowdhury and Amina Khatun, 2012,
―Image Compression Using Discrete Wavelet Transform‖, IJCSI
International Journal of Computer Science, Vol. 9- No 1.
[8] A.Alice Blessie, J. Nalini and S.C.Ramesh, 2011, ―Image
Compression Using Wavelet Transform Based on the Lifting
Scheme and its Implementation‖ IJCSI International Journal of
Computer Science Issues, Vol. 8-No. 1.
[9] Ms.Yamini S.Bute, Prof. R.W. Jasutkar, 2012, ―Implementation of
Discrete Wavelet Transform Processor For Image Compression‖
International Journal of Computer Science and Network (IJCSN),
(c) Vol. 1.
Fig 8. More on Residuals for wavelet 2D-compression [10] Radomir S. Stankovic, Bogdan J. Falkowski, 2003, ―The Haar
(a) Residuals (b) Histogram (c) Cumulative Histogram wavelet transform: its status and achievements‖ Computers and
Electrical Engineering, 25–44
V. CONCLUSION [11] P. Raviraj and M.Y. Sanavullah, 2007, ―The Modified 2D-Haar
Wavelet Transformation in Image Compression‖. Middle-East
Compression of image is an important field in Digital Journal of Scientific Research, 2 (2): 73-78.
signal processing. In this paper, comparison of various [12] Kamrul Hasan Talukder and Koichi Harada,2007, ―Haar Wavelet
transforms based image compression method is described. Based Approach for Image Compression and Quality Assessment of
If the energy retained is 100% then the compression is Compressed Image‖ Published in IAENG International Journal of
Applied Mathematics, 36:1.
known as lossless means we successfully restore all the
[13] Mohamed I. Mahmoud, Moawad I. M. Dessouky, Salah Deyab, and
information useful for user. This occurs when the threshold Fatma H. Elfouly. 2007, ―Comparison between Haar and
value is set to zero, meaning that the detail has not been Daubechies Wavelet Transformions on FPGA Technology‖
changed. If any values are changed then energy will be lost Proceedings Of World Academy Of Science, Engineering And
and this is known as lossy compression. Technology, VOLUME 20.

602

Potrebbero piacerti anche