Discrete wavelet transform using matlab

International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-

6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEME

252

DISCRETE WAVELET TRANSFORM USING MATLAB

Darshana Mistry1, Asim Banerjee

2

1

Asst. Professor, Computer Engineering, Indus Institute of Technology, Ahmedabad,

Gujarat, India 2Professor, Communication Technology, DAIICT, Gandhinagar, India

ABSTRACT:

In numerical analysis and functional analysis, a discrete wavelet transform (DWT) is

any wavelet transform for which the wavelets are discretely sampled. In this paper, there are

given fundamental of DWT and implementation in MATLAB. Image is filtered by low pass

(for smooth variation between gray level pixels) and high pass filter (for high variation

between gray level pixels). Image is decomposed into multilevel which include

approximation details (LL subband), horizontal detail (HL subband), vertical (LH subband)

and diagonal details (HH subband).

Keywords: Discrete Wavelet Transform (DWT), MATLAB, high pass filter, low pass filter.

I. INTRODUCTION

The transform of a signal is just another form of representing the signal. It does not

change the information content present in the signal.

Fourier Transmission (FT) representations do not include local information about the

original signals. Although the WFTs can provide localization information, they do not

provide flexible division of the time-frequency plane that can track slow changing

phenomena while providing more details for higher Frequencies. The wavelet representation

was introduced to correct the drawback of the former two methods using a multi-resolution

scheme.

The Wavelet Transform provides a time-frequency representation of the signal. A

wavelet series is representation of a square-integral (real or complex value) function by a

certain orthonormal (two vectors in an inner product space are orthonormal if they are

orthogonal (when two things can very independently or they are perpendicular) and all of unit

length).

INTERNATIONAL JOURNAL OF COMPUTER ENGINEERING

& TECHNOLOGY (IJCET)

ISSN 0976 – 6367(Print) ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), pp. 252-259 © IAEME: www.iaeme.com/ijcet.asp Journal Impact Factor (2013): 6.1302 (Calculated by GISI) www.jifactor.com

IJCET

© I A E M E



253

There are two classifications of wavelets [6]: (a) orthogonal (the low pass and high pass

filters have same length) and (b) biorthogonal (the low pass and high pass filters do not have

same length). Based on the application, either of them can be used.

The Wavelet transforms contribute to the desired sampling by filtering the signal with translations

and dilations of a basic function called “mother wavelet”. The mother wavelet can be used to

form orthonormal bases of wavelets, which is particularly useful for data reconstruction [2].

Daniel [1] represent the wavelet transform expands a signal, or function, into the wavelet

domain. As with any transform, like the Fourier or Gabor Transforms, the goal of expanding a

signal is to obtain information that is not apparent, or cannot be deduced, from the signal in its

original domain (usually space, time or distance).

A wavelet, in the sense of the Discrete Wavelet Transform (or DWT), is an orthogonal

function which can be applied to a finite group of data. Functionally, it is very much like the

Discrete Fourier Transform, in that the transforming function is orthogonal, a signal passed twice

through the transformation is unchanged, and the input signal is assumed to be a set of discrete-

time samples. Both transforms are convolutions.

Shripathi [6] introduce as The Discrete Wavelet Transform (DWT), which is based on sub-band

coding is found to yield a fast computation of Wavelet Transform. It is easy to implement and

reduces the computation time and resources required.

In DWT, the most prominent information in the signal appears in high amplitudes and the

less prominent information appears in very low amplitudes. Data compression can be achieved by

discarding these low amplitudes. The wavelet transforms enables high compression ratios with

good quality of reconstruction. Recently, the Wavelet Transforms have been chosen for the JPEG

2000 compression standard.

The discrete wavelet transform uses low-pass and high-pass filters, h(n) and g(n), to

expand a digital signal. They are referred to as analysis filters. The dilation performed for each

scale is now achieved by a decimator. The coefficients �� and �� are produced by convolving the

digital signal, with each filter, and then decimating the output. The �� coefficients are produced

by the low-pass filter, h(n), and called coarse coefficients. The �� coefficients are produced by

the high-pass filter and called detail coefficients. Coarse coefficients provide information about

low frequencies, and detail coefficients provide information about high frequencies. Coarse and

detail coefficients are produced at multiple scales by iterating the process on the coarse

coefficients of each scale. The entire process is computed using a tree-structured filter bank, as

seen in Fig. 1.

Fig. 1. Analysis filter bank. The high and low pass filters divide the signal into a series of coarse

and detail coefficients.

After analyzing, or processing, the signal in the wavelet domain it is often necessary to

return the signal back to its original domain. This is achieved using synthesis filters and

expanders. The wavelet coefficients are applied to a synthesis filter bank to restore the original

signal, as seen in Fig.2.



254

Fig. 2. Synthesis Filter Bank. The high and low pass filters combine the coefficients into the

original signal.

The discrete wavelet transform has a huge number of applications in science, engineering,

and mathematics and computer science. The wavelet domain representation of an image, or any

signal, is useful for many applications, such as compression, noise reduction, image registration,

watermarking, super-resolution etc.

II. TWO DIMENSIONAL DISCRETE WAVELET TRANSFORM

The wavelet transform can be expressed by the following equation (1):

F�a, b � f�xφ��,�∞

�∞�xdx ……….(1)

where the * is the complex conjugate symbol and function ψ is some function.

The discrete wavelet transform (DWT) is an implementation of the wavelet transform

using a discrete set of the wavelet scales and translations obeying some defined rules.

The two dimensional discrete wavelet transform is essentially a one dimensional analysis

of a two dimensional signal. It only operates on one dimension at a time, by analyzing the rows

and columns of an image in a separable fashion. The first step applies the analysis filters to the

rows of an image. This produces two new images, where one image is set or coarse row

coefficients, and the other a set of detail row coefficients. Next analysis filters are applied to the

columns of each new image, to produce four different images called sub bands. Rows and

columns analyzed with a high pass filter are designated with an H. Likewise, rows and columns

analyzed with a low pass filter are designated with an L. For example, if a subband image was

produced using a high pass filter on the rows and a low pass filter on the columns, it is called the

HL subband. Figure 3 shows this process in its entirety.

Fig. 3. Two Dimensional Discrete Wavelet Transform. The high and low pass filters operate

separable on the rows and columns to create four different subbands. An 8x8 image is used for

example purposes only.



255

Each subband provides different information about the image. The LL subband is a

coarse approximation of the image and removes all high frequency information. The LH

subband removes high frequency information along the rows and emphasizes high frequency

information along the columns. The result is an image in which vertical edges are

emphasized. The HL subband emphasizes horizontal edges, and the HH subband emphasizes

diagonal edges. To compute the DWT of the image at the next scale the process is applied

again to the LL subband (see fig. 4).

Each level of the wavelet decomposition, four new images are created from the

original N x N-pixel image. The size of these new images is reduced to ¼ of the original size,

i.e., the new size is N/2 x N/2. The new images are named according to the filter (low-pass or

high-pass) which is applied to the original image in horizontal and vertical directions. For

example, the LH image is a result of applying the low-pass filter in horizontal direction and

high-pass filter in vertical direction [2]. Thus, the four images produced from each

decomposition level are LL, LH, HL, and HH. The LL image is considered a reduced version

of the original as it retains most details. The LH image contains horizontal edge features,

while the HL contains vertical edge features. The HH contains high frequency information

only and is typically noisy and is, therefore, not useful for the registration. In wavelet

decomposition, only the LL image is used to produce the next level of decomposition (see

fig.5).

Fig.4. DWT image is based on approximate image detail (LL), horizontal details(HL),

vertical details(LH) and diagonal details(HH).

Fig. 5. Decomposed of and image level vise.



256

When we apply high frequency (use high pass filter) on an image, there are

high variations in the gray level between the two adjacent pixels. So edges are

occurred in image. When we apply low frequency (use low pass filter) on an image,

there are smooth variations between the adjacent pixels. So edges are not generated or

very few edges are generated. All information of image is remaining same as real

image information (it display as approximation image).

From fig. 6 and fig. 7 represent approximate details, horizontal details, vertical details

and diagonal details of and different images. Approximate details are same as original

image details. Horizontal details construct only horizontal information (edges).

Vertical details construct only vertical information (edges). Diagonal details construct

very few information of input image. So approximation image is applied into next

level for deformation.

(a)

(b)

Fig. 6. (a) Original image, (b) DWT image based on approximate image detail (LL),

horizontal details(HL), vertical details(LH) and diagonal details(HH) in one level.



257

(a)

(b)

Fig. 7. (a) original image, (b) DWT image based on approximate image detail (LL),

horizontal details(HL), vertical details(LH) and diagonal details(HH) in one level.

III. STEPS AND IMPLEMENTATION IN MATLAB

Basic steps are used to apply DWT in MATLAB (Matlab 2007b or later version).

1. Read an image.

2. Convert an input image into a gray scale image.

3. Perform a single-level wavelet decomposition(we get for information approximation,

horizontal, vertical and diagonal details of an image)

4. Construct and display approximations and details from the coefficients.

5. To display the results of the level 1 decomposition.

6. Regenerate an image by zero-level inverse Wavelet Transform.

7. Perform multilevel wavelet decomposition.

8. Extract approximation and detail coefficients. To extract the level 2 approximation

coefficients from step 5.

9. Reconstruct the Level 2 approximation and the Level 1 and 2 details.

10. Display the results of a multilevel decomposition.

11. Reconstruct the original image from the multilevel decomposition.

Fig. 8 (a) and 8(b) displayed images are decomposed into level 2 using DWT

algorithm.



258

IV. CONCLUSION

Using DWT, images are decomposing into four parts: Approximate image, Horizontal

details, Vertical details and diagonal details. When we apply high frequency on an image,

there are high variations in the gray level between the two adjacent pixels. So edges are

occurred in image. When we apply low frequency on an image, there are smooth variations

between the adjacent pixels. So edges are not generated or very few edges are generated. All

information of image is remaining same as real image information (it display as

approximation image).

(a)

(b)

Fig 8.(a) and (b) different images are decomposed up to second level

ACKNOWLEDGEMENT

I am very thankful to Dr Asim Banerjee who inspired and helped me to do this work.



259

REFERENCES

[1] Daniel L. Ward, “Redudant Discrete Wavelet Transform Based Super-Resolution Using

Sub-Pixel Image Registration”, thesis, March 2003.

[2] Prachya C., “High Performance Automatic Image Registration for Remote Sensing”, PhD

thesis, George Mason University, Fairfax, Virginia, 1999.

[3] Richa S., Mayank V., Afzel N., “Multimodal Medical Image Fusion using Redundant

Discrete Wavelet Transform”, International Conference on Advances in Pattern

Recognition, Feb. 2009.

[4] Milad Ghantous, Somik Ghosh, Magdy Bayoumi, “A multimodal automatic image

registration technique based on complex wavelets”, ICIP 2009, pp173-176.

[5] Time E., “Discrete Wavelet Transforms: Theory and Implementation”, Draft 2, June

1992.

[6] Shripathi D., “Discrete Wavelet Transform” , chapter -2, 2003.

[7] Bowman, M., Debray, S. K., and Peterson, L. L. 1993. Reasoning about naming systems.

[8] Ding, W. and Marchionini, G. 1997 A Study on Video Browsing Strategies. Technical

Report. University of Maryland at College Park.

[9] Fröhlich, B. and Plate, J. 2000. The cubic mouse: a new device for three-dimensional

input. In Proceedings of the SIGCHI Conference on Human Factors in Computing

Systems.

[10] Tavel, P. 2007 Modeling and Simulation Design. AK Peters Ltd.

[11] Sannella, M. J. 1994 Constraint Satisfaction and Debugging for Interactive User

Interfaces. Doctoral Thesis. UMI Order Number: UMI Order No. GAX95-09398.,

University of Washington.

[12] Forman, G. 2003. An extensive empirical study of feature selection metrics for text

classification. J. Mach. Learn. Res. 3 (Mar. 2003), 1289-1305.

[13] Brown, L. D., Hua, H., and Gao, C. 2003. A widget framework for augmented interaction

in SCAPE.

[14] Y.T. Yu, M.F. Lau, "A comparison of MC/DC, MUMCUT and several other coverage

criteria for logical decisions", Journal of Systems and Software, 2005, in press.

[15] Spector, A. Z. 1989. Achieving application requirements. In Distributed Systems, S.

Mullende.

[16] B.V. Santhosh Krishna, AL.Vallikannu, Punithavathy Mohan and E.S.Karthik Kumar,

“Satellite Image Classification Using Wavelet Transform”, International journal of

Electronics and Communication Engineering &Technology (IJECET), Volume 1, Issue 1,

2010, pp. 117 - 124, ISSN Print: 0976- 6464, ISSN Online: 0976 –6472.

[17] Dr. Sudeep D. Thepade and Mrs. Jyoti S.Kulkarni, “Novel Image Fusion Techniques

using Global and Local Kekre Wavelet Transforms” International journal of Computer

Engineering & Technology (IJCET), Volume 4, Issue 1, 2013, pp. 89 - 96, ISSN Print:

0976 – 6367, ISSN Online: 0976 – 6375.

[18] S. S. Tamboli and Dr. V. R. Udupi, “Compression Methods Using Wavelet Transform”

International journal of Computer Engineering & Technology (IJCET), Volume 3,

Issue 1, 2012, pp. 314 - 321, ISSN Print: 0976 – 6367, ISSN Online: 0976 – 6375.