(IJCSIS) International Journal of Computer Science and Information Security,

Vol. 9, No. 6, 2011

Instigation of Orthogonal Wavelet Transforms using

Walsh, Cosine, Hartley, Kekre Transforms and their
use in Image Compression
Dr. H. B.Kekre Dr. Tanuja K. Sarode Sudeep D. Thepade Ms. Sonal Shroff
Sr. Professor, Asst. Professor Associate Professor, Lecturer,
MPSTME, SVKM’s Thadomal Shahani Engg. MPSTME, SVKM’s Thadomal Shahani Engg.
NMIMS (Deemed-to-be College, NMIMS (Deemed-to-be College
University, Vileparle(W), Bandra (W), Mumbai-50, University, Vileparle(W), Bandra (W), Mumbai-50,
Mumbai-56, India. India. Mumbai-56, India. India.

Abstract—In this paper a novel orthogonal wavelet transform molecular dynamics, astrophysics, optics, quantum mechanics
generation method is proposed. To check the advantage of etc. This change has also occurred in image processing, blood-
wavelet transforms over the respective orthogonal transform in pressure, heart-rate and ECG analyses, DNA analysis, protein
image compression, the generated wavelet transforms are applied analysis, climatology, general signal processing, speech, face
to the color images of size 256x256x3 on each of the color planes
R, G, and B separately, and thus the transformed R, G, and B
recognition, computer graphics and multifractal analysis.
planes are obtained. Form each of these transformed color Wavelet transforms are also starting to be used for
planes, the 70% to 95% of the data (in form of coefficients having communication applications. One use of wavelet
lower energy values) is removed and image is reconstructed. The approximation is in data compression. Like other transforms,
orthogonal transforms Discrete Cosine Transform (DCT), Walsh wavelet transforms can be used to transform data then, encode
Transform, Hartley Transform and Kekre Transform are used the transformed data, resulting in effective compression [8].
for the generation of DCT Wavelets, Walsh Wavelets, Hartley Wavelet compression can be either lossless or lossy. The
Wavelets, and Kekre Wavelets respectively. From the results it is wavelet compression methods are adequate for representing
observed that the respective Wavelet transform outperforms the high-frequency components in two-dimensional images.
original orthogonal transform.
So far wavelets of only Haar transform have been studied.
The paper presents the wavelet generation of transforms alias,
I. INTRODUCTION Walsh transform, DCT, Hartley transform and Kekre
transform. Also the use of these transform wavelets is
The development of wavelets can be linked to several separate proposed and strudied for image compression. The
trains of thought, starting with Haar's work in the early 20th experimental results have shown better data compression can
century [16,17]. Wavelets are mathematical tools that can be be achieved in transform wavelets than using image
used to extract information from many different kinds of data, transforms themselves.
including images [21,22,24]. Sets of wavelets are generally
needed to analyze data fully. A set of "complementary" II. EXSISTING TRANSFORMS
wavelets will reconstruct data without gaps or overlap so that
the deconstruction process is mathematically reversible and is This section discusses some of the existing transforms, Walsh,
with minimal loss. Generally, wavelets are purposefully DCT, Hartley and Kekre.
crafted to have specific properties that make them useful for A. DCT
image processing. Wavelets can be combined, using a "shift,
A discrete cosine transform (DCT) expresses a sequence of
multiply and sum" technique called convolution, with portions
finitely many data points in terms of a sum of cosine functions
of an unknown signal(data) to extract information from the oscillating at different frequencies. In particular, a DCT is a
unknown signal. Wavelet transforms are now being adopted Fourier-related transform similar to the discrete Fourier
for a vast number of applications, often replacing the transform (DFT), but using only real numbers. DCTs are
conventional Fourier transform. They have advantages over equivalent to DFTs of roughly twice the length, operating on
traditional fourier methods in analyzing physical situations real data with even symmetry. There are eight standard DCT
where the signal contains discontinuities and sharp spikes[1- variants, of which four are common. The DCTs are important
4]. In fourier analysis the local properties of the signal are not to numerous applications in science and engineering, from
detected easily. STFT(Short Time Fourier Transform)[5] was lossy compression of audio and images to spectral methods for
introduced to overcome this difficulty. However it gives local the numerical solution of partial differential equations. For
properties at the cost of global properties. Wavelets overcome compression, the cosine functions are much more efficient
this shortcoming of Fourier analysis [6,7] as well as STFT. whereas for differential equations the cosines express a
Many areas of physics have seen this paradigm shift, including particular choice of boundary conditions.

125 http://sites.google.com/site/ijcsis/
ISSN 1947-5500
 (IJCSIS) International Journal of Computer Science and Information Security,
Vol. 9, No. 6, 2011
B. Walsh Transform
The Walsh matrix was proposed by Joseph Leonard Walsh ⎧ 1 ,x ≤ y
in 1923 [18,19]. Each row of a Walsh matrix corresponds to a ⎪
Walsh function. A Walsh matrix is a square matrix, with K x, y = ⎨− N + ( x + 1) , x = y + 1
dimensions a power of 2. The entries of the matrix are either +1 ⎪ 0 ,x > y +1
or −1. It has the property that the dot product of any two
distinct rows (or columns) is zero [20,23,25]. The sequency
⎩
ordering of the rows of the Walsh matrix can be derived from
(1)
the ordering of the Hadamard matrix by first applying the bit-
reversal permutation and then the Gray code permutation[9]. All diagonal elements and the upper diagonal elements are
The Walsh matrix (and Walsh functions) are used in computing one, while lower diagonal elements except the one exactly
the Walsh transform and have applications in the efficient below the diagonal are zero.
implementation of certain signal processing operations.
III. GENERATING WAVELET FROM ANY ORTHOGONAL
C. Hartley Transform TRANSFORM
Hartley transform was proposed by R. V. L. Hartley in Wavelet transform matrix of size P2 x P2 can be generated
1942, as an alternative to the Fourier transform[10]. It is one of from any orthogonal transform M of size PxP. For example, if
many known Fourier-related transforms. Compared to the we have orthogonal transform matrix of size 9x9, then its
Fourier transform, the Hartley transform has the advantages of corresponding wavelet transform matrix will have size 81x81.
transforming real functions to real functions (as opposed to i.e. for orthogonal matrix of size P, wavelet transform matrix
requiring complex numbers) and of being its own inverse. size will be Q, such that Q = P2.
D. Kekre Transform Consider orthogonal transform M of size pxp as shown
below.
Kekre transform[11] matrix is the generic version of
Kekre’s LUV color space matrix[12-15]. Most of the other
transform matrices have to be in powers of 2. This condition is
not required in Kekre transform. Any term in the Kekre
transform is generated as
M11 M12 ... M1 (P-1) M1P
M21 M22 ... M2 (P-1) M2P
. . ... . .
. . . .
MP1 MP2 ... MP (P-1) MPP
Figure 1 : PxP orthogonal transform matrix
1st column of M 2nd column of M pth column of M
Repeated P times Repeated P times Repeated P times
M11 M11 ... M11 M12 M12 ... M12 ... M1P M1P ... M1P

M21 M21 ... M21 M22 M22 ... M22 ... M2P M2P ... M2P

. . ... . . . ... . ... . . ... .

. . ... . . . ... . ... . . ... .
MP1 MP1 ... MP1 MP2 MP2 ... MP2 ... MPP MPP ... MPP
M21 M22 ... M2P 0 0 ... 0 ... 0 0 ... 0
0 0 ... 0 M21 M22 ... M2P ... 0 0 ... 0
. . . . . . . . . . . . .
. . . . . . . . . . . . .
0 0 ... 0 0 0 ... 0 ... M21 M22 ... M2P
... ... ...
... ... ...
MP1 MP2 ... MPP 0 0 ... 0 ... 0 0 ... 0
0 0 ... 0 MP1 MP2 ... MPP ... 0 0 ... 0
. . . . . . . . . . . . .
. . . . . . . . . . . . .
0 0 ... 0 0 0 ... 0 ... MP1 MP2 ... MPP
Figure.2: QxQ wavelet transform generated from PxP orthogonal transform( (Q = P2 )

126 http://sites.google.com/site/ijcsis/
ISSN 1947-5500
 (IJCSIS) International Journal of Computer Science and Information Security,
Vol. 9, No. 6, 2011
Figure 2 shows QxQ wavelet transform matrix generated from
PxP orthogonal transform matrix such that Q = P2. To Table 1,3,5 and 7 shows the comparison of MSE values
generate the wavelet matrix, the every column of the obtained from data compressed using DCT, Walsh, Hartley
orthogonal transform matrix is repeated P times. Then the and Kekre transforms applied on all the eleven test images
second row is translated P times to generate next P rows. respectively.
Similarly all rows are translated to generate P rows
corresponding to each row. Finally we get the wavelet matrix Table 2,4,6 and 8 shows the comparison of MSE values
of the size QxQ, where Q = P2 obtained from data compressed using DCT wavelet, Walsh
wavelet, Hartley wavelet and Kekre wavelet transforms
applied on all the eleven test images respectively.
IV. PROPOSED METHOD
In this section, the image compression using wavelet Figure 4: Comparison of average MSE with respect to 95% to
transform’s application is proposed. 70% of data compress using DCT wavelet, Walsh wavelet,
Step 1. Consider an image of size 256x256. The wavelet Hartley wavelet, Kekre wavelet, DCT, Walsh, Hartley and
transform matrix of size 256x256 is generated Kekre transform.
from orthogonal matrix of size 16x16.
Figure 5,6,7,8 shows the results of Balls image obtained from
Step 2. The wavelet transform is applied on each of the
DCT wavelet, Walsh wavelet, Hartley wavelet and Kekre
image plane i.e. R-plane, G-plane, B-plane wavelet respectively for 70% to 95% of data compress.
separately. Thus, transformed R-plane, G-plane,
B-plane are obtained.
Step 3. From the transformed R-plane, G-plane, B-plane
separately, the 70% to 95% coefficients having
lowest energy values are removed. And then the
image is reconstructed.
Step 4. Mean Square error between the reconstructed
image and the original image is computed.

V. RESULTS AND DISCUSSION

In this section, the image compression using wavelet
transform’s application is proposed. The proposed method is
implemented using MatLab 7.0 on Core 2 Duo processor.
DCT, Walsh, Hartley and Kekre wavelets were generated by
the method discussed in the section 3. The eleven different Figure 3:Eleven original color test images namely Aishwariya, Balls, Bird,
color images belonging to different categories, of size Boat, Flower, Ganesh, Scenary, Strawberry, Tajmahal, Tiger and Viharlake
256x256 were compressed using the proposed method. (from left to right and top to bottom) belonging to different categories

Figure 3 shows the eleven color test images of size 256x256x3

belonging to different categories.
Table 1: Comparison of MSE values obtained for 95% to 70% data compressed using DCT applied on all eleven images.
%data compressed 95 90 85 80 75 70
%data retained 5 10 15 20 25 30
Aishwariya 16.0803 8.1392 4.542 2.7457 1.7756 1.2044
Balls 75.5739 62.2747 50.7583 40.078 30.6593 22.6352
Bird 23.4414 19.7856 17.1511 14.846 12.6658 10.6395
Boat 63.2849 56.6238 49.9363 43.0231 36.116 29.7334
Flower 23.2196 13.3896 8.2092 5.158 3.3155 2.1642
Ganesh 66.9069 60.1663 53.5089 47.1965 41.0909 34.991
Scenary 32.4582 26.3064 21.5738 17.7571 14.5049 11.6967
Strawberry 42.358 30.1477 21.656 15.8716 11.6467 8.5902
Tajmahal 49.7616 39.457 30.907 23.5085 17.5614 12.7884
Tiger 67.5201 53.9452 42.909 33.8272 26.4423 20.1247
Viharlake 42.4999 35.5583 29.7518 24.3079 19.42 15.3702
Average 45.7368 36.89035 30.08213 24.39269 19.56349 15.4489

127 http://sites.google.com/site/ijcsis/
ISSN 1947-5500
 (IJCSIS) International Journal of Computer Science and Information Security,
Vol. 9, No. 6, 2011
Table 2: Comparison of MSE values obtained for 95% to 70% data compressed using DCT wavelets applied on all eleven images.
%data Compressed 95 90 85 80 75 70
%data retained 5 10 15 20 25 30
Aishwariya 13.9138 5.4578 2.5735 1.3751 0.7916 0.4853
Balls 67.3254 51.4344 38.2372 27.1908 18.4033 11.9991
Bird 15.1267 7.2307 3.5948 1.975 1.2121 0.7956
Boat 54.5218 44.1746 35.3552 27.2747 19.6998 13.4656
Flower 18.9017 7.4338 3.2424 1.5961 0.8935 0.5555
Ganesh 65.2921 56.5384 48.5545 40.7238 33.151 25.9447
Scenary 29.7195 20.5452 13.5051 8.2269 4.8059 2.7274
Strawberry 40.4291 27.01 17.4447 10.6523 6.2435 3.5483
Tajmahal 41.9902 29.0375 19.7188 12.8007 8.0717 5.0573
Tiger 65.7406 49.9408 37.7845 27.6822 19.5049 13.2931
Viharlake 38.3256 29.412 21.8512 15.5274 10.4845 6.8427
Average 41.02605 29.83775 21.98745 15.91136 11.20562 7.701327

Table 3: Comparison of MSE values obtained for 95% to 70% data compressed using Walsh transform applied on all eleven images.
%data Compressed 95 90 85 80 75 70
%data retained 5 10 15 20 25 30
Aishwariya 28.1012 18.5335 13.2878 9.6173 6.9535 4.9937
Balls 81.3445 72.1406 63.6927 55.4022 47.672 39.9922
Bird 29.0555 24.9949 21.8778 19.0824 16.3321 13.764
Boat 66.874 60.8236 55.0104 49.0244 42.8836 36.6938
Flower 36.1558 26.6554 20.6424 15.9753 12.2055 9.2632
Ganesh 71.3259 65.8241 60.7583 55.4545 49.8489 43.9028
Scenary 36.995 30.8505 26.1749 22.1124 18.3582 14.9854
Strawberry 50.8574 42.5104 35.597 29.5755 23.963 18.9102
Tajmahal 56.1151 46.753 39.0347 32.1825 26.1253 20.74
Tiger 76.8846 67.4124 59.7075 52.2118 44.8674 37.3638
Viharlake 46.1636 40.4746 35.3193 30.1336 25.1187 20.408
Average 52.71569 45.17936 39.19116 33.70654 28.57529 23.72883

Table 4: Comparison of MSE values obtained for 95% to 70% data compressed using Walsh wavelets applied on all eleven images.
%data compressed 95 90 85 80 75 70
%data retained 5 10 15 20 25 30
Aishwariya 24.2798 14.1835 8.4923 5.1467 3.2017 2.0188
Balls 71.2873 59.9048 50.0267 40.7423 31.9771 24.201
Bird 21.37 12.2772 6.7889 3.7529 2.1759 1.3296
Boat 57.1472 47.5557 39.4294 31.806 24.5572 17.8547
Flower 29.8175 18.8151 11.6577 6.9575 4.0902 2.3961
Ganesh 68.7866 61.5798 54.7242 47.9489 41.1375 34.1747
Scenary 33.2859 24.4968 17.4245 11.8686 7.8334 5.1252
Strawberry 47.5166 37.5645 29.5413 22.3788 16.4636 11.6458
Tajmahal 46.4039 34.44 25.3372 18.217 12.5737 8.4354
Tiger 74.0328 62.677 52.7802 43.7662 35.2198 27.521
Viharlake 41.6009 33.4509 26.2753 19.7348 14.0675 9.508
Average 46.86623 36.99503 29.31615 22.93815 17.57251 13.11003

Table 5: Comparison of MSE values obtained for 95% to 70% data compressed using Hartley transform applied on all eleven images.
%data compressed 95 90 85 80 75 70
%data retained 5 10 15 20 25 30
Aishwariya 17.5702 9.2244 5.3507 3.3036 2.144 1.4455
Balls 76.1777 62.8777 51.2288 40.6618 31.174 23.0542
Bird 24.0468 19.8922 17.1642 14.8753 12.6649 10.6047
Boat 63.743 56.9175 50.3089 43.4315 36.585 30.0655
Flower 23.245 13.3901 8.1816 5.132 3.3002 2.1621
Ganesh 67.4761 60.5399 54.0034 47.5902 41.4778 35.3603
Scenary 33.3664 26.7334 22.0444 18.2306 14.8923 11.9873
Strawberry 43.5429 31.7365 23.165 17.109 12.8251 9.5586
Tajmahal 49.833 39.5094 30.807 23.4858 17.5716 12.8477
Tiger 67.9142 54.7827 44.2022 35.0442 27.6414 21.3417
Viharlake 43.0531 35.9083 30.1151 24.6871 19.7647 15.5462
Average 46.36076 37.41019 30.59739 24.86828 20.00373 15.8158

128 http://sites.google.com/site/ijcsis/
ISSN 1947-5500
 (IJCSIS) International Journal of Computer Science and Information Security,
Vol. 9, No. 6, 2011
Table 6: Comparison of MSE values obtained for 95% to 70% data compressed using Hartley wavelets applied on all eleven images.
%data compressed 95 90 85 80 75 70
%data retained 5 10 15 20 25 30
Aishwariya 25.213 13.1629 7.0521 4.0416 2.4411 1.5258
Balls 71.1273 57.2092 45.1996 34.8136 25.8726 18.5723
Bird 23.6101 13.4562 7.3995 4.0684 2.3198 1.4062
Boat 57.524 47.4436 38.7197 30.4678 22.8547 16.2068
Flower 30.4708 17.2874 9.5996 5.2322 2.9442 1.6909
Ganesh 68.6207 59.8072 51.8952 44.4754 36.9332 29.7399
Scenary 33.7757 24.0033 16.4779 10.8463 6.9661 4.4146
Strawberry 47.5595 35.0583 25.241 17.6 12.0036 8.0276
Tajmahal 45.9356 33.1194 23.4711 16.1064 10.6663 6.9495
Tiger 71.4263 57.522 46.3082 36.2968 27.6191 20.1696
Viharlake 40.7793 31.5152 23.8878 17.1808 11.82 7.7949
Average 46.91294 35.41679 26.84106 20.10266 14.76734 10.59074

Table 7: Comparison of MSE values obtained for 95% to 70% data compressed using Kekre transform applied on all eleven images.
%data compressed 95 90 85 80 75 70
%data retained 5 10 15 20 25 30
Aishwariya 104.7137 98.3782 89.9261 79.8792 71.1881 63.4727
Balls 95.4663 94.6395 91.7457 85.9287 78.2736 70.4866
Bird 71.0262 66.9461 61.3682 53.14 44.6532 36.7741
Boat 96.431 91.469 85.2311 77.9438 70.1285 62.199
Flower 75.3232 73.098 70.8086 66.3523 61.4276 54.4408
Ganesh 89.5643 85.6352 79.994 73.6673 66.8255 59.6006
Scenary 69.4835 66.4225 60.9814 54.9584 48.8438 42.7133
Strawberry 91.5023 86.9626 82.2165 76.2722 69.2823 61.9791
Tajmahal 87.7596 81.797 74.3466 66.3947 58.3154 50.4829
Tiger 103.1722 96.723 90.382 84.1157 77.8424 71.2272
Viharlake 65.3572 60.0596 54.1362 48.2416 42.1389 36.1324
Average 86.34541 82.01188 76.46695 69.71763 62.62903 55.40988

Table 8: Comparison of MSE values obtained for 95% to 70% data compressed using Kekre wavelets applied on all eleven images.
%data compressed 95 90 85 80 75 70
%data retained 5 10 15 20 25 30
Aishwariya 41.1364 30.8323 22.567 15.7857 10.5192 6.8614
Balls 78.2216 69.5023 60.7559 51.6477 42.7972 34.186
Bird 30.9089 18.9243 11.0784 6.3362 3.7573 2.3177
Boat 61.2107 51.5525 43.1938 35.063 27.1555 20.0594
Flower 43.864 33.5713 24.2075 16.0388 9.7355 5.4288
Ganesh 73.9496 67.8945 61.3198 54.3184 46.8956 39.3169
Scenary 40.3419 30.1718 22.2297 15.6323 10.7463 7.3131
Strawberry 58.15 49.6929 41.7717 34.4112 27.2761 20.7789
Tajmahal 53.9875 41.6384 32.1871 23.958 16.8148 11.1398
Tiger 86.4872 76.9422 68.3485 60.3057 51.9975 43.7026
Viharlake 46.5487 39.6689 32.3033 25.1071 18.4207 12.622
Average 55.8915 46.39922 38.17843 30.78219 24.19234 18.5206

Table 9: Comparison of average MSE values obtained for 95% to 70% data compressed using DCT, Walsh, Hartley, Kekre transforms and their
corresponding wavelets applied on all eleven images.

%data compressed 95 90 85 80 75 70
%data retained 5 10 15 20 25 30
DCT Wavelets 41.02605 29.83775 21.98745 15.91136 11.20562 7.701327
Walsh Wavelets 46.86623 36.99503 29.31615 22.93815 17.57251 13.11003
Hartley Wavelets 46.91294 35.41679 26.84106 20.10266 14.76734 10.59074
Kekre Wavelets 55.8915 46.39922 38.17843 30.78219 24.19234 18.5206
DCT 45.7368 36.89035 30.08213 24.39269 19.56349 15.4489
Walsh 52.71569 45.17936 39.19116 33.70654 28.57529 23.72883
Hartley 46.36076 37.41019 30.59739 24.86828 20.00373 15.8158
Kekre 86.34541 82.01188 76.46695 69.71763 62.62903 55.40988

129 http://sites.google.com/site/ijcsis/
ISSN 1947-5500
 (IJCSIS) International Journal of Computer Science and Information Security,
Vol. 9, No. 6, 2011

Figure 4: Comparison of average MSE with respect to 95% to 70% of data compressed using DCT wavelet, Walsh wavelet, Hartley wavelet, Kekre wavelet,
DCT, Walsh, Hartley and Kekre transform.

Figure 5: Results of Balls image obtained from DCT wavelet for 70% to 95% of data compressed.

130 http://sites.google.com/site/ijcsis/
ISSN 1947-5500
 (IJCSIS) International Journal of Computer Science and Information Security,
Vol. 9, No. 6, 2011

Figure 6: Results of Balls image obtained from Walsh wavelet for 70% to 95% of data compressed.

Figure 7: Results of Balls image obtained from Hartley wavelet for 70% to 95% of data compressed.

131 http://sites.google.com/site/ijcsis/
ISSN 1947-5500
 (IJCSIS) International Journal of Computer Science and Information Security,
Vol. 9, No. 6, 2011

Figure 8: Results of Balls image obtained from Kekre wavelet for 70% to 95% of data compressed.

From Table 9, it is observed that performance of all Proceedings of the International Computer Music Conference (ICMC-
87, Tokyo), Computer Music Association, 1987.
wavelet transforms is better than that of their respective
[6] S. Mallat, "A Theory of Multiresolution Signal Decomposition: The
orthogonal transforms as indicated by lower MSE values. Wavelet Representation," IEEE Trans. Pattern Analysis and Machine
Figure 4 compares the average MSE with respect to 95% Intelligence, vol. 11, pp. 674-693, 1989.
to 70% od data compress using. DCT wavelet, Walsh [7] Strang G. "Wavelet Transforms Versus Fourier Transforms." Bull.
wavelet, Hartley wavelet, Kekre wavelet, DCT, Walsh, Amer. Math. Soc. 28, 288-305, 1993.
Hartley and Kekre transform. [8] P. P. Kanjilal, “Adaptive Prediction and Predictive Control”, IET, p 210,
1995
VI. CONCLUSION [9] Yuen, C. “Remarks on the Ordering of Walsh Functions”, IEEE
Transactions on Computers, C-21: 1452, 1972.
In this paper, novel orthogonal wavelet transform
[10] Hartley, R. V. L., “A more symmetrical Fourier analysis applied to
generation method is proposed. The proposed method can transmission problems”, Proc. IRE 30, 144–150, 1942.
be used to generate the wavelet transform from any [11] H.B.Kekre, Sudeep D. Thepade, “Image Retrieval using Non-
orthogonal transform. To test the efficiency of wavelet Involutional Orthogonal Kekre’s Transform”, International Journal of
transform, they are applied on the eleven different color Multidisciplinary Research and Advances in Engineering (IJMRAE),
Ascent Publication House, 2009, Volume 1, No.I, 2009. Abstract
images for the purpose of data compression. The available online at www.ascent-journals.com
orthogonal transforms used in this paper are DCT, Walsh, [12] H. B.Kekre, Sudeep D. Thepade, “Image Blending in Vista Creation
Hartley and Kekre. From the results, it can be concluded using Kekre's LUV Color Space”, SPIT-IEEE Colloquium and Int.
that wavelet transforms outperforms their respective Conference, SPIT, Andheri, Mumbai, 04-05 Feb 2008.
orthogonal transform as indicated by lower MSE values [13] H.B.Kekre, Sudeep D. Thepade, “Boosting Block Truncation Coding
using Kekre’s LUV Color Space for Image Retrieval”, WASET Int.
REFERENCES Journal of Electrical, Computer and System Engineering (IJECSE),
Vol.2, Num.3, Summer 2008. Available online at
www.waset.org/ijecse/v2/v2-3-23.pdf
[1] K. P. Soman and K.I. Ramachandran. ”Insight into WAVELETS From [14] H.B.Kekre, Sudeep D. Thepade, “Color Traits Transfer to Grayscale
Theory to Practice”, Printice -Hall India, pp 3-7, 2005. Images”, In Proc.of IEEE First International Conference on Emerging
[2] Raghuveer M. Rao and Ajit S. Bopardika. “Wavelet Transforms – Trends in Engg. & Technology, (ICETET-08), G.H.Raisoni COE,
Introduction to Theory and Applications”, Addison Wesley Longman, Nagpur, INDIA. Available on IEEE Xplore.
pp 1-20, 1998. [15] H.B.Kekre, Sudeep D. Thepade, “Creating the Color Panoramic
[3] C.S. Burrus, R.A. Gopinath, and H. Guo. “Introduction to Wavelets and Viewusing Medley of Grayscale and Color Partial Images”, WASET Int.
Wavelet Transform” Prentice-hall International, Inc., New Jersey, 1998. Journal of Electrical, Computer and System Engg. (IJECSE), Volume 2,
[4] Amara Graps, ”An Introduction to Wavelets”, IEEE Computational No. 3, Summer 2008. Available online at www.waset.org/ijecse/v2/v2-3-
Science and Engineering, vol. 2, num. 2, Summer 1995, USA. 26.pdf
[5] Julius O. Smith III and Xavier SerraP“, An Analysis/Synthesis Program [16] Dr. H.B.kekre, Sudeep D. Thepade, Adib Parkar, “A Comparison of
for Non-Harmonic Sounds Based on a Sinusoidal Representation'', Haar Wavelets and Kekre’s Wavelets for Storing Colour Information in
a Greyscale Image”, International Journal of Computer Applications

132 http://sites.google.com/site/ijcsis/
ISSN 1947-5500
 (IJCSIS) International Journal of Computer Science and Information Security,
Vol. 9, No. 6, 2011
(IJCA), Volume 1, Number 11, December 2010, pp 32-38. Available at Professor at MPSTME, SVKM’s NMIMS. He has guided 17 Ph.Ds, more than
www.ijcaonline.org/archives/volume11/number11/1625-2186 100 M.E./M.Tech and several B.E./ B.Tech projects. His areas of interest are
[17] Dr. H.B.kekre, Sudeep D. Thepade, Adib Parkar “Storage of Colour Digital Signal processing, Image Processing and Computer Networking. He
Information in a Greyscale Image using Haar Wavelets and Various has more than 270 papers in National / International Conferences and Journals
Colour Spaces”, International Journal of Computer Applications (IJCA), to his credit. He was Senior Member of IEEE. Presently He is Fellow of IETE
Volume 6, Number 7, pp.18-24, September 2010. Available online at and Life Member of ISTE Recently 11 students working under his guidance
http://www.ijcaonline.org/volume6/number7/pxc3871421.pdf have received best paper awards. Two of his students have been awarded Ph.
[18] Dr.H.B.Kekre, Sudeep D. Thepade, Juhi Jain, Naman Agrawal, “IRIS D. from NMIMS University. Currently he is guiding ten Ph.D. students.
Recognition using Texture Features Extracted from Walshlet Pyramid”,
ACM-International Conference and Workshop on Emerging Trends in Dr. Tanuja K. Sarode has Received Bsc.(Mathematics) from Mumbai
Technology (ICWET 2011),Thakur College of Engg. And Tech., University in 1996, Bsc.Tech.(Computer
Mumbai, 26-27 Feb 2011. Also will be uploaded on online ACM Portal. Technology) from Mumbai University in 1999,
M.E. (Computer Engineering) degree from
[19] Dr.H.B.Kekre, Sudeep D. Thepade, Akshay Maloo, “Face Recognition
Mumbai University in 2004, Ph.D. from Mukesh
using Texture Features Extracted form Walshlet Pyramid”, ACEEE
Patel School of Technology, Management and
International Journal on Recent Trends in Engineering and Technology
Engineering, SVKM’s NMIMS University, Vile-
(IJRTET), Volume 5, Issue 1, www.searchdl.org/journal/IJRTET2010
Parle (W), Mumbai, INDIA. She has more than 12
[20] Dr.H.B.Kekre, Sudeep D. Thepade, Juhi Jain, Naman Agrawal, years of experience in teaching. Currently working
“Performance Comparison of IRIS Recognition Techniques using as Assistant Professor in Dept. of Computer
Wavelet Pyramids of Walsh, Haar and Kekre Wavelet Transforms”, Engineering at Thadomal Shahani Engineering
International Journal of Computer Applications (IJCA), Number 2, College, Mumbai. She is life member of IETE, member of International
Article 4, March 2011, Association of Engineers (IAENG) and International Association of Computer
http://www.ijcaonline.org/proceedings/icwet/number2/2070-aca386 Science and Information Technology (IACSIT), Singapore. Her areas of
[21] Dr.H.B.Kekre, Sudeep D. Thepade, Akshay Maloo, “Face Recognition interest are Image Processing, Signal Processing and Computer Graphics. She
using Texture Features Extracted from Haarlet Pyramid”, International has 90 papers in National /International Conferences/journal to her credit.
Journal of Computer Applications (IJCA), Volume 12, Number 5,
December 2010, pp 41-45. Available at Sudeep D. Thepade has Received B.E.(Computer) degree from North
www.ijcaonline.org/archives/volume12/number5/1672-2256 Maharashtra University with Distinction in
[22] Dr.H.B.Kekre, Sudeep D. Thepade, Juhi Jain, Naman Agrawal, “IRIS 2003. M.E. in Computer Engineering from
Recognition using Texture Features Extracted from Haarlet Pyramid”, University of Mumbai in 2008 with
International Journal of Computer Applications (IJCA), Volume 11, Distinction, currently submitted thesis for
Number 12, December 2010, pp 1-5, Available at Ph.D. at SVKM’s NMIMS, Mumbai. He has
www.ijcaonline.org/archives/volume11/number12/1638-2202. more than 08 years of experience in teaching
[23] Dr.H.B.Kekre, Sudeep D. Thepade, Akshay Maloo, “Performance and industry. He was Lecturer in Dept. of
Comparison of Image Retrieval Techniques using Wavelet Pyramids of Information Technology at Thadomal Shahani
Walsh, Haar and Kekre Transforms”, International Journal of Computer Engineering College, Bandra(w), Mumbai for
Applications (IJCA) Volume 4, Number 10, August 2010 Edition, pp 1- nearly 04 years. Currently working as
8, http://www.ijcaonline.org/archives/volume4/number10/866-1216 Associate Professor in Computer Engineering
at Mukesh Patel School of Technology
[24] Dr.H.B.Kekre, Sudeep D. Thepade, Akshay Maloo, “Query by image Management and Engineering, SVKM’s NMIMS, Vile Parle(w), Mumbai,
content using color texture features extracted from Haar wavelet INDIA. He is member of International Association of Engineers (IAENG) and
pyramid”, International Journal of Computer Applications (IJCA) for the International Association of Computer Science and Information Technology
special edition on “Computer Aided Soft Computing Techniques for (IACSIT), Singapore. He is member of International Advisory Committee for
Imaging and Biomedical Applications”, Number 2, Article 2, August many International Conferences. He is reviewer for various International
2010. http://www.ijcaonline.org/specialissues/casct/number2/1006-41 Journals. His areas of interest are Image Processing Applications, Biometric
[25] Dr.H.B.Kekre, Sudeep D. Thepade, “Image Retrieval using Color- Identification. He has about 110 papers in National/International
Texture Features Extracted from Walshlet Pyramid”, ICGST Conferences/Journals to his credit with a Best Paper Award at International
International Journal on Graphics, Vision and Image Processing (GVIP), Conference SSPCCIN-2008, Second Best Paper Award at ThinkQuest-2009
Volume 10, Issue I, Feb.2010, pp.9-18, Available online National Level paper presentation competition for faculty, Best paper award at
www.icgst.com/gvip/Volume10/Issue1/P1150938876.html Springer international conference ICCCT-2010 and second best research
project award at ‘Manshodhan-2010’.
AUTHORS PROFILE
Dr. H. B. Kekre has received B.E. (Hons.) in Telecomm. Engineering. from Ms. Sonal Shroff has Received B.Sc.(Physics) from University of Mumbai in
Jabalpur University in 1958, M.Tech 1996, B.Sc.Tech.(Computer Technology) from
(Industrial Electronics) from IIT Bombay in University of Mumbai in 1999. She has more than
1960, M.S.Engg. (Electrical Engg.) from 10 years of experience in teaching. Currently
University of Ottawa in 1965 and Ph.D. working as Lecturer in Dept. of Computer
(System Identification) from IIT Bombay Engineering at Thadomal Shahani Engineering
in 1970 He has worked as Faculty of College. She is life member of ISTE. Her areas of
Electrical Engg. and then HOD Computer interest are Image Processing, Signal Processing
Science and Engg. at IIT Bombay. For 13 and Computer Graphics.
years he was working as a professor and head
in the Department of Computer Engg. at
Thadomal Shahani Engineering. College, Mumbai. Now he is Senior

133 http://sites.google.com/site/ijcsis/
ISSN 1947-5500